Dexmate co-founders Tao Chen, Yuzhe Qin, and Chongyang Wang are developing general-purpose humanoid robots at a pace unthinkable only a few years ago.
Launched in 2024, the NVIDIA Inception startup’s Vega humanoid, available to order, offers a Swiss Army Knife of capabilities: arms that lift, fingers that grasp, torsos that fold and more.
Dexmate is targeting customers in manufacturing, logistics and retail. Its units are already in testing with public companies, said Chen.
“Both me and my co-founders have full-stack experience. We want to build the entire robot, and we feel like we can build something better than the robots currently on the market,” said Chen, Dexmate’s CEO.
The market for humanoid robots could exceed $5 trillion by 2050, with some 1 billion of them in such areas as supply chains, networks for repair, maintenance and support, according to Morgan Stanley.
Dexmate’s swift time to market is no coincidence. During his undergraduate coursework, Chen began his lifelong ambition of building physical robots. As he progressed in his graduate and post-graduate pursuits, his ability to develop more capable intelligence for physical robots accelerated along with the advances from NVIDIA GPUs and Isaac Sim.
“We have a long history of using NVIDIA tools. During our PhDs we used Isaac Gym, the first generation of GPU-based simulation, and later Isaac Sim and Isaac Lab — we have been using it for five or six years, and now in our company heavily,” said Chen.
Dexmate recently showed its Vega humanoid robot at the NVIDIA Santa Clara headquarters, making a splash by effortlessly lifting barbells.
“When we have a customer request to automate tasks, we go to their factory, we do the scanning, and we create a digital twin and run simulations in Isaac Sim with synthetic data,” said Chen. “And the VLA models we are using are based on NVIDIA Isaac GR00T.”
Dexmate is ready for the next chapter. Its Vega runs on Orin AGX, but the company is building physical AI applications that rely on multimodal foundation models and can integrate vision, language and reasoning.
“In the next couple of years, the AI models are going to be much better,” said Chen. “We will need much better compute to do the edge onboard inference for these models.”
- Security Camera Installation – indoor/outdoor IP CCTV systems & video analytics
- Access Control Installation – key card, fob, biometric & cloud‑based door entry
- Business Security Systems – integrated alarms, surveillance & access control
- Structured Cabling Services – voice, data & fiber infrastructure for new or existing builds
- Video Monitoring Services – 24/7 remote surveillance and analytics monitoring
Author: 360 Technology Group
