SAN JOSE, Calif. — NVIDIA today announced the latest member of its Jetson family of artificial intelligence platforms – it’s not Elroy or Rosie, but rather something that could help developers create a Rosie in real life.
The Jetson Nano is an AI computer that delivers 472 GFLOPS of computing performance for running modern AI workloads, with power consumption as little as 5 watts. Unveiled at today’s GPU Technology Conference by NVIDIA founder and CEO Jensen Huang, Jetson Nano will come in two versions – a $99 development kit (available now) aimed at developers, makers, and enthusiasts; and a $129 production-ready module (shipping June 2019) for companies looking to create mass-market edge systems.
Jetson Nano supports high-resolution sensors, can process many sensors in parallel, and run multiple modern neural networks on each sensor stream, the company said in its announcement. It also supports many AI frameworks, giving developers options for integrating their preferred models and frameworks into the product.
“Jetson Nano makes AI more accessible to everyone – and is supported by the same underlying architecture and software that powers our nation’s supercomputers,” said Deepu Talla, vice president and general manager of autonomous machines at NVIDIA. “Bringing AI to the maker movement opens up a whole new world of innovation, inspiring people to create the next big thing.”
Reaching the makers
The Nano platform joins the Jetson family of systems, including the Jetson AGX Xavier, designed for autonomous machines; and Jetson TX2, designed for AI at the edge. With Nano, NVIDIA said the Jetson system can extend its reach to more than 30 million makers, developers, inventors and students.
NVIDIA said the power of AI is largely out of reach for this market, because “typical technologies do not pack enough computing power and lack an AI software platform.” The $99 price point helps bring AI technologies to this space, allowing them to build AI projects that weren’t previously possible, and take existing projects to the next level, including mobile robots, drones, digital assistants, and automated appliances.
The developer kit will provide out-of-the-box support for full desktop Linux, compatibility with peripherals and accessories, and ready-to-use projects and tutorials, NVIDIA said. Users will also be able to access the Jetson developer forum to ask technical questions of other developers.
“The Jetson Nano Developer Kit is exciting because it brings advanced AI to the DIY movement in a really easy-to-use way,” said Chris Anderson of DIY Robocars, DIY Drones, and the Linux Foundation’s Dronecode project. “We’re planning to introduce this technology to our maker communities because it’s a powerful, fun, and affordable platform that’s a great way to teach deep learning and robotics to a broader audience.”
As an additional incentive to the maker market, NVIDIA said it created a reference platform to “jumpstart the building of AI applications” by minimizing the time spent on hardware assembly. The NVIDIA JetBot is a small mobile robot that can be built with off-the-shelf components, and will be open-sourced on GitHub.
Jetson module for mass-market companies
The $129 module (in quantities of 1,000 or more) is aimed at providing companies with new options for embedded applications, including network video recorders, home robots, and intelligent gateways with full analytics capabilities, NVIDIA said. By reducing the time spent in hardware design, testing, and verification of the AI system, users of the Nano Module will be able to bring products to market faster, the company added.
“Cisco Collaboration is on a mission to connect everyone, everywhere for rich and immersive meetings,” said Sandeep Mehra, vice president and general manager for Webex Devices at Cisco. “Our work with NVIDIA and use of the Jetson family lineup is key to our success. We’re able to drive new experiences that enable people to work better, thanks to the Jetson platform’s advanced AI at the edge capabilities.”
In order to help customers move their AI and machine learning workloads to the edge, NVIDIA announced working with Amazon Web Services to qualify its AWS Internet of Things Greengrass to run optimally with Jetson-powered devices.
System specs and hardware
Key features of Jetson Nano module and developer kit include:
- GPU: 128-core NVIDIA Maxwell architecture-based GPU
- CPU: Quad-core ARM A57
- Video: 4K resolution at 30 frames per second (H.264/H.265), 4K at 60 fps (H.264/H.265) encode and decode.
- Camera: CSI-2 DPHY lanes, 12x (module) and 1x (developer kit)
- Memory: 4 GB 64-bit LPDDR4; 25.6 GB/sec.
- Connectivity: Gigabit Ethernet
- OS Support: Linux for Tegra
- Module size: 70mm x 45 mm
- Developer Kit size: 100mm x 80mm