NVIDIA has launched the Jetson AGX Thor, a next-generation supercomputer designed for physical AI and robotics. Embedded computing specialist, Syslogic is developing a rugged AI computer based on the new module.
The NVIDIA Jetson Thor modules deliver up to 2,070 FP4 TFLOPS of AI performance – 7.5 times that of the NVIDIA AGX Orin – while achieving 3.5 times greater energy efficiency. Power consumption can be configured between 40 and 130 watts.
Edge computing redefined
With these capabilities, the NVIDIA Jetson Thor series sets a new standard for high-performance Edge computing. Data from multiple high-resolution sensors can be processed in milliseconds. Potential applications include real-time AI reasoning, autonomous navigation and multimodal AI agents for humanoid robotics, autonomous transport vehicles and AMRs (autonomous mobile robots). Image processing, sensor fusion (LiDAR, radar and cameras) and decision-making can be carried out virtually without latency.
Engineered for extreme environments
Syslogic ensures that NVIDIA technology can be deployed even in the harshest environments. Syslogic has previously used Jetson modules as the basis for its robust AI computers, which are used worldwide in sectors such as agricultural technology, construction machinery, mining, AMRs and railways. These robust embedded systems handle inferencing and intelligent vision tasks wherever extreme reliability is required. The new Rugged Computer RML A5AGX, which is based on the Jetson Thor module, will be available by the end of 2025. It is designed to operate within extended temperature ranges and complies with IP67/IP69 protection standards.
Michael Jung, Product Manager at Syslogic, comments: “With the Jetson AGX Thor, NVIDIA is taking Edge computing to a whole new level.” Jung adds that Syslogic is excited to be part of this technological leap forward.
