ASUS IoT has launched yet another new industrial-grade hardware in the form of the PE3000N system.
This new model is based on the latest NVIDIA Jetson Thor platform equipped with Blackwell GPU, a 14-core Arm CPU, and up to 128GB of LPDDR5X memory that work together to provide up to 2,070 FP4 TFLOPS of AI performance while maintaining a small, power-efficient footprint, making it ideal for robotics and automation systems where space and energy efficiency matter.
It is built for extreme environments of operating optimally from -20°C to 60°C, supports up to 4x 25GbE links and 16x GMSL cameras for high-bandwidth sensor fusion, alongside a 12–60V DC input and ignition support for stable operation across industrial, automotive, and smart-city applications.
The modular design allows for expansion via PoE, GMSL, CAN, and QSFP28 interfaces while also capable of being vertically expanded within a compact 2U form factor and includes PTP/PPS for precise synchronization, TPM 2.0 security, and LTE/5G/GNSS module support.
All of these are designed to address demanding AI workloads on the edge, such as generative AI, visual language models (VLMs), and large language models (LLMs), while enabling real-time video analytics and autonomous control. Since this is based on NVIDIA’s hardware, of course, it works seamlessly with Team Green’s ecosystem, including Isaac for robotics, Holoscan for real-time sensor processing, and Blueprint for Video Search and Summarization from NVIDIA Metropolis, all processed locally.
Availability
ASUS IoT plans to officially launch the PE3000N in Q1 2026, with sample inquiries open through local representatives by Q4 2025.
