Nvidia has launched a new addition to its Jetson product line: a credit card-sized (70x45mm) form factor delivering up to 21 trillion operations/second (TOPS) of throughput, according to the company.

The Jetson Xavier NX module consumes as little as 10 watts of power, costs $399 and is designed to be embedded in edge devices, such as small robots, drones, sensors for factory logistics and production lines, optical inspection, portable medical devices and other industrial IoT systems that demand small, light, low-power processing.

“AI has become the enabling technology for modern robotics and embedded devices that will transform industries,” said Deepu Talla, VP/GM of Edge Computing at Nvidia. “Many of these devices, based on small form factors and lower power, were constrained from adding more AI features. Jetson Xavier NX lets our customers and partners dramatically increase AI capabilities without increasing the size or power consumption of the device.”

At 10W of power, the Jetson Xavier NX delivers up to 14 TOPS, according to Nvidia, and at 15W up to 21 TOPS. At the high range of its power, said Rob Csongor, Nvidia VP, Autonomous Machines, in a pre-launch press briefing, the system delivers the performance at the lower range of the previously announced Jetson Xavier but in a nano form factor. “We’ve taken Xavier and tuned it for extremely high performance but within a very targeted power budget,” he said.

Jetson Xavier NX can run multiple neural networks in parallel while processing data from multiple high-resolution sensors simultaneously, Nvidia said. The hardware runs on the CUDA-X AI software architecture used in existing Jetson offerings and is supported by Nvidia’s JetPack software development kit, an AI software stack that the company said “can run modern and complex AI networks, accelerated libraries for deep learning as well as computer vision, computer graphics, multimedia and more.”

“Nvidia’s embedded Jetson products have been accelerating the research, development and deployment of embedded AI solutions on Lockheed Martin’s platforms,” said Lee Ritholtz, director and chief architect of applied AI at Lockheed Martin. “With Jetson Xavier NX’s exceptional performance, small form factor and low power, we will be able to do more processing in real time at the edge than ever before.”

Nvidia also announced today that it finished first in five MLPerf Inference 0.5 benchmarks measuring the performance of AI inference workloads. Nvidia said the results demonstrate the inference capabilities of the company’s Turing GPUs for data centers and the Jetson Xavier system-on-a-chip for edge. The Jetson Xavier NX module is built around a low-power version of the Xavier SoC used in these benchmarks.

Specifications for the Jetson Xavier NX include the Nvidia Volta GPU with 384 CUDA cores and 48 Tensor cores; 6-core Carmel Arm 64-bit CPU; up to six CSI cameras (36 via virtual channels); 8GB 128-bit LPDDR4x memory at 51.2GB/second using Gigabit Ethernet; and Ubuntu-based Linux operating system.

Jetson Xavier NX joins the Jetson product line that includes Jetson Nano, and the Jetson AGX Xavier and Jetson TX2 series. The company said Jetson offering support major AI frameworks, including TensorFlow, PyTorch, MxNet, Caffe and others. The Jetson Xavier NX will be available in March, but Nvidia said developers can begin now using the Jetson AGX Xavier Developer Kit with a software patch to emulate Jetson Xavier NX.

“In a world where AI chips are announced on what seems like a daily basis, I believe Nvidia raised the bar with its Jetson Xavier NX — showing that exceptional performance at small size and low power, together with a consistent and powerful software architecture, is what matters in embedded edge computing,” said Patrick Moorhead, president and principal analyst of Moor Insights & Strategy.

This article originally appeared on EnterpriseAI.