Nvidia Launches Jetson Orin Nano Super: Doubled AI Performance at Half the Price

NVIDIA recently launched the Jetson Orin Nano Super Developer Kit, a device designed to provide powerful processing capabilities for artificial intelligence and robotics projects, catering to both hobbyists and professional developers. The new version is priced at $249, which is half the cost of previous models, and offers significantly enhanced AI processing performance.

This mini-computer can perform up to 67 trillion operations per second (TOPS), representing a 70% increase in neural network processing power compared to its predecessor. The memory bandwidth has also been increased by 50%, reaching 102GB/s, which accelerates data transmission and processing efficiency. These improvements are made possible by a new power mode that boosts the clock speeds of the GPU, memory, and CPU.

Based on the same hardware architecture as the original Orin Nano, the Jetson Orin Nano Super includes a reference carrier board and an 8GB system module with an NVIDIA Ampere GPU (featuring tensor cores) and a six-core Arm CPU. Existing Orin Nano users can expect similar performance enhancements through an upcoming JetPack update.

The device supports complex generative AI models, large language models, and robotics applications, making it suitable for tasks such as building chatbots or visual AI agents. Despite its compact size and power consumption ranging from 7W to 25W, it can run advanced AI applications typically requiring larger systems.

For current Jetson Orin Nano users, NVIDIA will provide software updates to enhance generative AI inference performance. Additionally, the device is compatible with NVIDIA's AI ecosystem, including the Isaac platform for robotics and the Metropolis platform for visual AI, facilitating the development and deployment of AI applications.

Overall, the Jetson Orin Nano Super makes generative AI capabilities more accessible, enabling the creation of solutions in areas like robotics and computer vision. The lower price point and improved performance make this device a compelling choice for edge computing applications.