In an era where artificial intelligence is rapidly moving into the physical world, embodied intelligence is emerging as a core driving force behind transformation in fields such as industrial manufacturing, healthcare rehabilitation, and smart logistics. Embodied intelligence not only requires AI to possess powerful perception, reasoning, and decision-making capabilities, but also demands low-latency, high-reliability computing support at the edge to meet the stringent environmental and performance requirements of industrial scenarios.

As a NVIDIA NPN Elite Partner, Pinli Technology has always been committed to deeply integrating cutting-edge AI computing technologies with real-world application scenarios.

Today, Pinli Technology officially launches an embodied intelligence computing platform based on the latest NVIDIA Jetson Thor™, the industrial-grade development kit Y-C28-DEV. It is designed to bring powerful computing capabilities and development efficiency to embodied intelligence scenarios, accelerating the transition from prototype validation to large-scale deployment.

  Powerful Edge AI Computing Support  

NVIDIA Jetson Thor™ is a computing platform purpose-built for Physical AI and humanoid robots. It features a next-generation GPU architecture and CPU, integrates a Transformer engine and high-performance multimodal perception capabilities, and delivers up to 2070 FP4 TFLOPS of AI compute performance.

The Y-C28-DEV supports real-time inference of various generative AI models and complex control algorithms at the edge. It can process multi-source data from depth cameras, RGB cameras, LiDAR, millimeter-wave radar, and IMU sensors, enabling key capabilities such as environmental perception, path planning, human-robot interaction, and autonomous decision-making. It meets the real-time requirements of industrial robots, AGVs/AMRs, robotic arms, and production line inspection scenarios.

  Full-Scenario Interface Coverage  

The Y-C28-DEV development kit provides rich I/O interfaces to support expansion across embodied intelligence scenarios.

The system supports up to 8K video output, compatible with industrial monitoring and digital twin visualization applications. MIPI CSI supports up to 16 synchronized camera inputs, meeting high-concurrency data acquisition requirements such as machine vision and AGV navigation. USB interfaces support high-resolution industrial cameras and legacy industrial equipment.

In addition, it supports expansion of 5G and Wi-Fi 6 wireless communication modules to enable real-time data transmission in mobile scenarios. Flexible communication protocol expansion is supported, compatible with fieldbus control systems, enabling seamless integration with existing production lines and cloud platforms for edge-cloud collaboration.

  Industrial-Grade Reliability and Stability  

The Y-C28-DEV adopts a compact structural design with dimensions of only 120mm × 140mm × 65mm, enabling flexible integration into industrial equipment and robotic systems. The cooling fan is specifically designed for high-power operation. The product meets industrial-grade standards, ensuring stable and continuous operation in harsh environments while significantly reducing maintenance costs and downtime risks. It has a lifecycle of 5–10 years and comes with a 3-year warranty.

  Full-Stack Support to Accelerate AI Application Development  

The system comes pre-installed with an operating system and is ready to use out of the box. It supports NVIDIA® Isaac platform and GR00T model framework, and is compatible with toolchains such as CUDA, Metropolis, and Holoscan, providing end-to-end support from model training, optimization, deployment to real-time inference. This helps developers rapidly build safe and scalable embodied intelligence solutions.

Pinli Technology’s industrial-grade embodied intelligence computing platform represents not only a leap forward in edge AI computing power, but also a foundational enabler that empowers robots with stronger environmental perception, understanding, and interaction capabilities—injecting powerful momentum into the transition of artificial intelligence from the digital world to the physical world.