Despite the advancements of artificial intelligence, not many people can get their hands on such expensive tech. Computers and cloud systems powering A.I. with powers such as instantaneous translation, image, and verbal recognition, and stronger decision-making capabilities can cost businesses quite the pretty penny.
The Jetson Nano is Nvidia’s response to this issue of pricey A.I.
Like its bigger brothers in the Jetson production line, the Nano is an embedded computing board that can be slapped onto just about any A.I.-applicable device. Since the computer is directly attached to the device, latency and security issues plaguing remote devices aren’t as common; making the machine process information faster with less chance of being hacked.
Measuring 69.6 x 45mm, the Jetson Nano is the smallest Jetson device yet. The A.I. computer runs on 5-10 watts of power and has 472 gigaflops for running A.I. algorithms. The GPU uses 128 Nvidia CUDA cores while the CPU uses a Quad-core ARM Cortex A57 MPCore processor. It supports A.I. programming software such as PyTorch, TensorFlow, and Keras, and comes with your standard USB 3.0 and 2.0 Micro-B ports, Gigabit Ethernet, and a microSD storage slot.
While the $99 developer kit version of the Jetson Nano is aimed specifically for designers, DIY makers, and researchers, a production-ready version costing $129 is also going to be available for companies who are too busy to program their own A.I. Nvidia hopes this lower price point gets more developers to dive into the world of A.I. without worrying too much about costs and barrier of entry. Expect our robot overlords to rise up when the Jetson Nano becomes available this coming June 2019.
Learn more over at Nvidia.