- NVIDIA introduces $249 Jetson Orin Nano Super with 67 TOPS AI processing power.
- New device operates at 25 watts while matching RTX 2060 performance for AI tasks.
- Memory bandwidth increased to 102GB/s, supporting simultaneous processing from 4 cameras.
- Device can run AI models with up to 8 billion parameters, including Llama 3.1.
- Software updates available for existing Jetson Orin Nano owners to improve efficiency.
Nvidia’s latest edge computing device brings enterprise-grade AI capabilities to the mass market, with the launch of the Jetson Orin Nano Super at $249, half the price of its predecessor while delivering 70% better performance.
Technical Specifications and Performance
The palm-sized computing unit achieves 67 TOPS (trillion operations per second) for AI processing, utilizing 1,024 CUDA cores and consuming between 7-25 watts of power. This represents significant energy efficiency compared to traditional gaming GPUs that require 115-170 watts.
The device features:
- 6-core ARM processor
- 8GB shared memory
- 102GB/s memory bandwidth
- Ampere architecture GPU
Performance metrics indicate the Jetson Orin Nano Super matches the AI processing capabilities of an RTX 2060 while consuming one-seventh of the power. This efficiency makes it particularly suitable for edge computing applications like robotics, drone control, and smart camera systems.
AI Model Support and Applications
The device supports various language models, including those with 8 billion parameters. When running quantized versions of models like Llama 3.1, it generates approximately 18-20 tokens per second.
Practical applications include:
- Local AI chatbot deployment
- Multi-camera feed processing
- Robotics control systems
- Stable Diffusion image generation
Nvidia CEO Jensen Huang stated: "This is a brand new Jetson Nano Super. Almost 70 trillion operations per second, 25 watts and $249. It runs everything the HGX does, it even runs LLMs."
The company has also released software updates to enhance the performance of existing Jetson Orin Nano devices, demonstrating ongoing support for their edge computing ecosystem.
While the device shows promise for prototyping and small-scale applications, its 8GB memory limitation may restrict usage in large-scale enterprise deployments requiring more extensive computational resources.
✅ Follow BITNEWSBOT on Facebook, LinkedIn, X.com, and Google News for instant updates.
Previous Articles:
- Trump Plans U.S. Bitcoin Strategic Reserve, Sending Crypto Past $100,000
- Bitcoin Hits Record $108K Before Fed-Driven Pullback to $92K
- Trump Allies Push Bitcoin Reserves Plan as Three US States Join Crypto Treasury Race
- Bitcoin Plunges to $90K as Fed’s Powell Warns of Sticky Inflation
- Tether Makes $775M Bet on Free-Speech Video Platform Rumble, Stock Surges 41%
