Specs-wise, the new Jetson AGX Xavier module is design to provide AI Server performance level, but at relatively low power consumptions of 10W, 15W, and 30W. It also comes built with 512 CUDA Tensor core, an 8-cores CPU, and up to 32 Deep Learning TOPS. NVIDIA aims to leverage its new platform towards institutions and scientific bodies that have invested heavily in to researching and advancement of AI, deep learning, neural networks, and machine learning. On that note, its also supports currently existing SDKs, such as JetPack and DeepStream.
Early current users and partners of NVIDIA’s new Jetson AGX Xavier platform include the likes of DJI and Lockheed Martin (for their robotics programs), all the way to Nanopore Technologies. A company that specialises in DNA sequencing.
The Jetson AGX Xavier platform is already available in the US, and retails for US$1099 (~RM4597). Again the module is designed for enthusiasts interested in developing their own AI-driven autonomous systems, or the likes thereof. At the time of writing, NVIDIA did not state if the platform will be made available to institutes in Malaysia.
