NTT Docomo Business launches MEC×GPUaaS for physical AI

The announcement sits inside a wider NTT Group push. NTT Docomo Business presents its services as a realization of its AI infrastructure concept, the AI-Centric ICT Platform, which it positions at the core of the group's next-generation communications vision, IOWN. Against that backdrop, Kazuo Onishi described the company's physical AI approach as offering, as a service, the functions robots would rather offload: the latest security features, advanced video AI processing, and wide-area management of large numbers of devices.

The two announced services tackle different layers of the same problem. MEC×GPUaaS addresses data sovereignty, safety and scalability. The company links those to concrete risks: leakage of personal or confidential data, accidents in the real world caused by AI misjudgment, and missed opportunities from failing to scale. It then maps its features onto them — completion inside its carrier network, a closed network with security functions, and on-demand dynamic resource use. The company also points to road-vehicle cooperative autonomous driving as a physical AI example, gathering road-surface, oncoming vehicle, pedestrian and blind-spot information from cars and smart poles, and notes that physical AI and sovereign AI are closely related.

The Pro menu targets a different pain: connected products spreading across vehicles, industrial machinery, surveillance cameras and medical devices, with one view cited that global IoT connections could reach 39 billion by 2030. Managing line activation, suspension and cancellation, linking external systems, tracking communications and handling faults gets complicated, and cyberattacks on connected products and their networks raise the need for security across the communication path. Onishi was asked whether the company would lead ecosystem-building across fields; he said he wants to, but will first support each field's own moves. How far that support turns into genuine cross-field leadership — and how deeply such projects reach into individual industries — looks like the thing to watch.