XPENG is developing robotics-related technologies as part of a broader strategy to apply artificial intelligence beyond passenger vehicles. The company’s Vision-Language-Action (VLA 2.0) architecture, originally designed for automated driving, is also intended to support other physical AI systems such as humanoid robots and robotaxi fleets.
The company describes the AI model underlying VLA 2.0 as a foundation that can be deployed across different types of machines operating in real-world environments. In addition to vehicles, XPENG is exploring applications including humanoid robots and modular flying vehicles. These initiatives form part of what the company describes as a broader physical AI strategy focused on developing a shared artificial intelligence platform capable of controlling intelligent machines operating in real-world environments.
Within the near term, XPENG is focusing on robotaxi testing using vehicles equipped with the VLA 2.0 system. These vehicles are undergoing supervised road testing in China, with pilot operations scheduled to begin later in the year. The robotaxi program serves as one of the first operational environments for the AI architecture and is intended to support the development of higher levels of automated driving.
The company also confirmed that Volkswagen will become the first commercial partner to adopt the VLA 2.0 intelligent driving system in vehicles sold in China. The collaboration reflects a growing pattern of cooperation between established automotive manufacturers and technology-focused companies in the development of software-defined vehicles and advanced driver assistance systems.
XPENG states that the same AI framework can potentially be adapted for robotics and other autonomous platforms that interact with physical environments, extending its use beyond traditional automotive applications.
