NVIDIA (NASDAQ: NVDA) and T-Mobile (NASDAQ: TMUS) announced a collaboration to deploy physical AI applications over distributed edge networks using AI-RAN infrastructure. The partnership involves Nokia and multiple developers to transform wireless networks into platforms for edge AI computing.

T-Mobile is piloting NVIDIA RTX PRO 6000 Blackwell Server Edition AI infrastructure to demonstrate physical AI applications at network edge locations. The initiative builds on T-Mobile's position as the first U.S. carrier to pilot NVIDIA's AI-RAN infrastructure with Nokia's anyRAN software.

Physical AI developers including Fogsphere, LinkerVision, Levatas, Vaidio and Siemens Energy are building reasoning and vision AI agents using the NVIDIA Metropolis Blueprint for video search and summarization. The City of San Jose is among the first to assess the technology for smart city operations.

NVIDIA introduced version 3 of its VSS blueprint, which features agentic information retrieval capabilities that can search video footage for specific events in under five seconds. The blueprint includes a modular architecture and can summarize long-form video up to 100 times faster than manual reviews, according to the company.

The collaboration addresses latency and connectivity requirements for physical AI applications that need to operate in real-time across urban intersections, industrial facilities and rural areas. The architecture enables AI processing to be offloaded from individual devices to nearby edge computing locations.

Pilot applications include traffic optimization systems being tested with the City of San Jose, automated utility line inspections, facility management systems, and industrial safety monitoring for hazardous environments. Partners using the VSS blueprint include Caterpillar, KION, Hitachi, HCLTech, and Tulip.