PATEO Integrates NVIDIA and ByteDance AI Models into Automotive Cabin Systems
PATEO has introduced an in-vehicle AI solution using NVIDIA DRIVE AGX Thor and the Doubao language model to enable offline voice interaction and automated cabin management.
PATEO is deploying NVIDIA Nemotron and ByteDance’s 35-billion-parameter Doubao language model within its new AI Box, a hardware unit built on the NVIDIA DRIVE AGX Thor platform. The integration allows vehicles to perform local AI inference, enabling voice interactions and automated cabin functions to operate without a constant cloud connection.
The system acts as an agent that combines voice, visual, and sensor data to interpret user requests and monitor cabin conditions. According to the company, the platform can manage climate settings, provide safety alerts, and suggest adjustments based on driving habits and environmental data. Because the system performs compute tasks locally, it maintains functionality in areas with poor network coverage.
PATEO has secured a project nomination from an undisclosed new energy vehicle manufacturer, with mass production scheduled for the second quarter of 2027. The company is positioning this technology as a bridge to recurring revenue models, moving away from traditional one-time hardware sales toward ongoing AI service operations.
Beyond the automotive sector, PATEO is exploring the application of this platform in robotics. The company is also working to enable software updates for existing vehicles, allowing manufacturers to add AI capabilities to older hardware through PATEO's modular software architecture.