Nvidia Introduces Affordable $250 'Jetson Computer' for Local AI Experimentation by Enthusiasts
Nvidia Introduces Affordable $250 'Jetson Computer' for Local AI Experimentation by Enthusiasts
Nvidia has launched a new variant of its Jetson computer, priced at $249, designed for running local artificial intelligence applications. The pocket-sized Orin Nano is claimed to boost speed and efficiency by twice while costing half as much as its predecessor. The device can handle approximately 70% more computational tasks, according to Nvidia.
The Orin Nano is a great option for DIY enthusiasts looking to develop their own AI applications or for developers of robots and industrial tools to execute complex applications without connecting to the cloud.
During a YouTube video unveiling the product, Nvidia CEO Jensen Huang showcases the small, palm-sized computer in his kitchen. Huang explains that the computer can handle nearly "seventy trillion" operations per second and consumes just 25 watts of power.
Nvidia CEO Jensen Huang presented Nvidia's latest Jetpack computer for local AI programming on Tuesday. Image credit: Nvidia
As Huang explains, "A long time ago, we devised a unique type of processor, which we named a robotics processor. No one grasped our intentions back then, but we envisioned that these deep learning models would advance, and we'd soon have robots for everything." The Jetpack computers were initially designed for robotics, but they can now also run sophisticated AI models like Meta's Llama.
Sam Altman's company World (formerly Worldcoin), which utilizes iris scans for human ID verification, incorporates a Jetpack module within its Orb scanning device. According to a post published in October, World mentioned that the "newest iteration" of the Orb features the most advanced NVIDIA Jetson module with nearly 5 times the AI performance over the previous version, enabling faster and smoother human verification checks.
In essence, the Orin Nano acts as a portable brain that can be integrated with other hardware to activate its AI capabilities. Cloud service providers like Amazon and Google charge for server and AI model access, and these costs can be substantial. Applications that demand guaranteed uptime and minimal latency, such as warehouse robots, may struggle with remote cloud hosting. In such scenarios, a device like the Orin Nano, capable of handling lightweight AI applications, would be a better fit. However, it should be noted that the Orin Nano cannot replace Nvidia's high-end GPUs, which cost tens of thousands of dollars and have the capability to train and infer from large-scale AI models.
If you're a hobbyist aiming to create a new AI-equipped robot, the Orin Nano might be a suitable choice.
The Orin Nano's affordability and improved capabilities, thanks to artificial-intelligence and advanced technology, make it an attractive option for hobbyists interested in developing AI-powered robots. In the future, with continued advancements in tech, devices like the Orin Nano could revolutionize the way we integrate AI into various industries and applications.