Nvidia has announced the Jetson AGX Orin, an AI accelerator aimed at robotics projects, autonomous machines, medical devices and AI edge computing. Based on the Ampere architecture, and with additional Arm Cortex A78AE CPUs, it offers six times the computing power of its predecessor, the Jetson AGX Xavier.
Powered by a 12 core Arm A78AE, 32GB of LPDDR5 and Ampere, The Jetson AGX Orin targets the large, complex projects such as models for natural language processing, 3D perception, and fusing the inputs from multiple sensors. Designed to be used at the edge rather than at server level, the Orin is capable of two trillion operations a second. At full load the Jetson AGX Orin consumes only 50W of power.
“As robotics and embedded computing transform manufacturing, healthcare, retail, transportation, smart cities and other essential sectors of the economy, the demand for processing continues to surge,” said Deepu Talla, vice president and general manager of embedded and edge computing at Nvidia. “Jetson AGX Orin addresses this need, enabling the 850,000 Jetson developers and over 6,000 companies building commercial products on it to create and deploy autonomous machines and edge AI applications that once seemed impossible.”
The company also announced the release of its Omniverse Replicator platform, used for training AIs through synthetic data generation and not, as we first thought, producing hot earl grey tea. Although presumably you could train an AI to do just that.
At just 100 x 87mm, the Jetson Orin board retains the small form factor for which the line is known. Both the module and its developer kit will be available in Q1 2022, and you can sign up at the web page to be notified when it is.