AI corporations DEEPX linked Amazon Web Services (AWS) cloud and Internet of Things (IoT) infrastructure with its own Neural Processing Unit (NPU) to build an environment that can remotely deploy and manage AI models to robots and industrial equipment. Based on this, it will expand its push into the U.S. Robotics and industrial vision and smart factory markets.
DEEPX said on the 1st that AWS disclosed on its official tech blog on the 27th that it combined the DEEPX "DX-M1" NPU with "AWS IoT Core" and "AWS IoT Greengrass" to deploy and manage AI models and NPU runtime environments to multiple edge devices.
An NPU is a semiconductor specialized for AI model inference operations. In Edge AI, it is used to perform AI computation on on-site devices such as cameras, robots, and industrial PCs without sending data to the cloud for processing.
By using the "DEEPX Greengrass Solution" offered on AWS Marketplace, ONNX models, the open AI model standard format, can be converted into the DXNN format that can run on DEEPX NPUs.
NPU drivers and firmware, as well as the runtime software needed to run AI models on actual equipment, can also be deployed remotely using AWS IoT Greengrass. The company said the same AI runtime environment can be applied or models and software can be updated without visiting equipment installed in factories, logistics centers, stores, or robots.
DEEPX plans to expand a structure in which AI model development and management are carried out in the cloud, while low-power NPUs in the field handle actual inference. The company described this as a "center-edge federated AI" structure, in which the cloud and the edge split roles to develop, deploy, and operate AI.
DEEPX CEO Kim Nok-won also attended the "AWS Startup Partner Summit 2026" held in San Francisco from Jan. 31 to Feb. 1. DEEPX plans to leverage AWS's global partner ecosystem to expand this technical integration into overseas customer projects, including in the United States.
First, it will push to apply the DX-M1-based deployment environment to U.S. Robotics and industrial vision and smart factory clients. It is also reviewing a plan to expand the application scope to the next-generation Generative AI Semiconductor "DX-M2."
A DEEPX official said, "For physical AI to spread in earnest, not only low-power AI Semiconductor but also infrastructure that can easily deploy AI developed in the cloud to countless field devices and manage it continuously is needed," and added, "We will broaden the foundation for physical AI to spread to real industrial sites by consolidation of AWS's cloud and IoT ecosystem with DEEPX's ultra-low-power AI Semiconductor technology."