Cities Worldwide Build and Scale AI Services With NVIDIA Physical and Agentic AI

NVIDIA says cities worldwide are adopting its physical and agentic AI stack to turn camera, sensor, and operational data into actionable services, building on initiatives announced at GTC Berlin. Shared computing infrastructure lets city departments, neighboring municipalities, and local developers build on a common foundation without each acquiring resources of its own. Depending on the workload, NVIDIA says teams can start with RTX workstations or a small GPU system rather than a major infrastructure investment.

In Spain, 1MillionBot combines open source ALIA models trained at the Barcelona Supercomputing Center with NVIDIA Nemotron in resident-service agents for Madrid Digital (Comunidad de Madrid) and Administració Oberta de Catalunya. The company reports 150+ deployments across Spain handling municipal procedures, appointment requests, urban-planning incidents, and service inquiries. In the San Francisco Bay Area, LiveX is using NVIDIA Cosmos world models, Nemotron open models, and Omniverse and Metropolis libraries on RTX PRO GPUs for conversational AI hosts that can be tailored to local languages and service needs in airports, hospitals, and public events.

Vision agents are also being applied to existing camera feeds. In Wrocław, Poland, K2K is building shared AI infrastructure using Cosmos world reasoning models and the Metropolis Blueprint for video search and summarization to convert camera streams into intelligence for mobility, safety, and event response — reusing the same infrastructure across several city functions. In Malaysia, ITMAX bundles infrastructure and agentic AI into long-term managed services, with SamurAI V2 vision agents scaling on RTX PRO 6000 GPUs to support 17,000+ cameras by 2027. ITMAX also integrates TeraSim, an open source traffic simulation platform from Laplace Intelligence, to test signal timing; the company says this reduces peak-hour corridor travel times by up to 26%.

Beyond public-facing services, digital twins are being positioned for infrastructure planning and operational simulation. At Dangote's Lekki refinery in Nigeria, Honeywell Technologies' Forge Plant Cognition uses Nemotron-powered agents to interpret operational data, with an Omniverse-based digital twin of the plant planned for simulation and training. The platform runs on Dell PowerEdge servers with NVIDIA RTX PRO accelerators and helps operators spot bottlenecks and anticipate equipment and process issues.

Open questions center on scale and governance: whether the reported gains, such as ITMAX's corridor travel-time reductions, hold across varied city conditions, and how local governments will manage sensitive data processing as camera-based agents expand. Watch also for how quickly shared-infrastructure models spread beyond the partner deployments named here, and whether smaller municipalities adopt the workstation-scale entry point NVIDIA describes.

keyFacts:[