| Processing the latest AI workloads at 472 GFLOPS
Achieves high power efficiency with 5W power consumption
Jetson Nano Developer Kit, Just $99

NVIDIA announced the Jetson Nano, an AI computer that enables the construction of millions of intelligent systems, on the 19th.
The Jetson Nano, a small AI computer based on CUDA-X, processes the latest AI workloads at a computational speed of 472 GFLOPS (gigaflops) and is highly power-efficient with a power consumption of only 5W.

NVIDIA founder and CEO Jensen Huang unveiled two versions of the Jetson Nano at the GPU Technology Conference (GTC 2019): a $99 developer kit for developers and manufacturers, and a production-ready module for enterprises looking to build general-purpose edge systems.

The Jetson Nano supports high-resolution sensors and can process multiple sensors simultaneously, as well as run multiple modern neural networks on each sensor stream. It supports multiple popular AI frameworks simultaneously, allowing developers to easily integrate their preferred models or frameworks into their products.

The Jetson Nano is included in the Jetson family of products, along with the Jetson AGX Xavier for autonomous machines and the Jetson TX2 for edge AI. With the launch of the Jetson Nano, the Jetson product family, which is suitable for companies, startups, and researchers, is now available to over 30 million manufacturers, developers, inventors, and students worldwide.

Deepu Talla, Executive Vice President and General Manager of Autonomous Machines at NVIDIA, stated, “We expect the Jetson Nano to significantly enhance AI accessibility for everyone,” adding, “The Jetson Nano supports the same architecture and software used in supercomputers.”

He added, “As AI is applied to open-source manufacturing, a new world of innovation will open up,” and “the Jetson Nano will provide new inspiration for the development of next-generation innovative products.”


Jetson Nano Developer Kit
Due to insufficient computing power and a lack of AI software platforms, the influence of AI is largely limited in the manufacturing community and the education sector.
The Jetson Nano developer kit, priced at $99, is expected to bring a new wave of innovation to manufacturers, inventors, developers, and students alike by enabling access to the latest AI capabilities on an affordable platform. This enables the construction of AI projects that were previously impossible, and allows existing projects such as mobile robots, drones, digital assistants, and automated devices to be upgraded to the next level.

The Jetson Nano Developer Kit supports Linux for desktop and is compatible with a wide range of popular peripherals and accessories. In addition, it comes with RTU projects and tutorials that enable manufacturers to quickly and easily launch AI projects. NVIDIA also provides answers to technical questions for users through the Jetson Developer Forum.

Chris Anderson of DIY Robocars, DIY Drones, and the Linux Foundation’s Dronecode Project stated, “The reason the Jetson Nano developer kit is interesting is that it allows users to easily use AI while integrating cutting-edge AI into the DIY movement,” adding, “We plan to introduce this technology to our community, which can teach deep learning and robotics to more people.”


Jetson Nano Module
In the past, companies struggled with issues regarding size, power, cost, and AI computational density. The Jetson Nano module addresses the requirements of embedded applications, including network video recorders, home robots, and intelligent gateways. This not only reduces overall development time but also accelerates time-to-market by shortening the time spent on hardware design, testing, and verification of complex, powerful, and power-efficient AI systems.

The Jetson Nano module also comes with power management, clocking, memory, and fully accessible input/output. Because AI workloads are fully software-defined, enterprises can update performance and capabilities even after system deployment.

Sandeep Mehra, Vice President and General Manager of Cisco Webex Device, said, “Cisco Collaboration aims to connect everyone and enable rich, immersive meetings anywhere,” adding, “Collaboration with NVIDIA and the use of the Jetson product family are key to realizing this vision.”

NVIDIA is collaborating with Amazon Web Services (AWS) to enable customers to easily move AI and machine learning workloads to the edge. Through this, AWS IoT Greengrass is optimized to run on Jetson-based devices such as the Jetson Nano.

“Our customers span a wide range of industries, including energy management, industrial, logistics, smart buildings, and smart homes,” said Dirk Didascalou, Vice President of AWS IoT. “They are building intelligence and computer vision into their apps to use services at the edge in real time.”

He added, “Through AWS IoT GreenGrass, customers can perform local inference on Jetson-based devices and resend relevant data to the cloud to improve model training.”


The entire Jetson line using a single software
NVIDIA CUDA-X is a collection of more than 40 acceleration libraries that enable modern computing apps to take advantage of NVIDIA's GPU-accelerated computing platform. The JetPack SDK is AI software built on CUDA-X and based on acceleration libraries for deep learning, computer vision, computer graphics, and multimedia processing that support the entire Jetson line.

Jetpack comes with the latest versions of CUDA, cuDNN, TensorRT, and a full desktop Linux OS. Jetson is compatible with the NVIDIA AI platform.


Reference platform for rapid prototype design
NVIDIA has established a reference platform that enables the rapid start of building AI applications by minimizing the time required for initial hardware assembly. The NVIDIA JetBot is a small mobile robot that can be built using off-the-shelf parts or open source on GitHub.
The NVIDIA Jetson Nano developer kit is currently selling for $99. The Jetson Nano module is priced at $129 (for orders of 1,000 or more) with delivery scheduled for June. Both products are available for purchase through NVIDIA's global distributors. The developer kit is also available at Seeed Studio and SparkFun.