Insider Brief
- NEC’s corporate venture capital fund invested an undisclosed amount in Dexmate, a U.S. robotics startup developing the Vega wheeled humanoid robot and an integrated physical AI development platform.
- Vega runs AI models onboard using an Nvidia Jetson Thor with 128 GB of memory and about 2,070 AI TFLOPS, while its 4.8 kWh battery provides about 20 hours of runtime or more than 16 hours at full AI compute load.
- NEC said it plans to explore physical AI applications with Dexmate across multiple industries, combining the startup’s robotics platform with NEC’s technology and industry experience.
NEC’s corporate venture capital fund announced it has invested in Dexmate, a U.S. robotics startup developing the Vega humanoid robot and a physical AI development platform.
The investment was made through the NEC Orchestrating Future Fund, or NOFF, according to NEC. The amount invested was not disclosed. Dexmate was founded in 2024 by AI robotics researchers and is developing an integrated system for AI-powered robots.
NEC pointed out that physical AI systems require a range of technologies beyond AI models, including sensors, control systems, training data, simulation environments and on-site tuning. Dexmate indicated it is developing its platform to bring those elements together and streamline the process of building robots for specific tasks and deploying them in operating environments.
“Dexmate’s humanoid robot “VEGA” and integrated Physical AI development platform fundamentally lower the barriers to robot development and deployment by enabling anyone to easily design, configure, and operate AI-powered robots in real-world environments,” Dexmate co-Founder and CEO Tao Chen said in the announcement. “Through this strategic partnership with NEC and their global expertise, we look forward to deploying our technology at scale and jointly accelerating the real-world implementation of physical AI across a wide range of industries.”
What is Dexmate’s Vega
Vega is a wheeled humanoid robot that runs AI models onboard using an Nvidia Jetson Thor with 128 GB of memory and about 2,070 AI TFLOPS of computing performance, according to Dexmate. Its 4.8 kWh battery provides about 20 hours of runtime, or more than 16 hours with its AI compute operating at full load.
According to the company, the robot has more than two meters of reach across its workspace, with arms that can each carry 15 lbs. and are about 15 centimeters longer than a human. It’s three-section folding torso is capable of reaching from floor level to more than two meters high. Dexmate also uses hardware-synchronized sensors intended to produce aligned data that can be used for training AI models.
Vega’s safety systems include brakes on major joints, real-time self-collision checking, a validation layer that blocks commands outside defined limits and a software-accessible emergency stop.
NEC said it plans to explore collaboration with Dexmate on physical AI applications across multiple industries, combining the startup’s robotics technology with NEC’s existing technologies and industry experience.
“NEC has long been engaged in the research, development, and real-world implementation of AI technologies,” added NEC Corporate SVP Shigeki Wada. “Through dialogue with Dexmate, we aim to explore new possibilities for value creation in the Physical AI domain and, together with our partners, contribute to solving the challenges faced by customers and society.”
Image credit: Dexmate
