NVIDIA Partners Deploy Physical AI Across Global Cities at Smart City Expo World Congress

By 2050, two-thirds of the world's population will live in cities, adding 2.5 billion people to urban areas. This surge demands smarter infrastructure, with the smart traffic management market projected to hit $20 billion by 2027. At this week's Smart City Expo World Congress in Barcelona, NVIDIA's partners are demonstrating how physical AI—combining digital twins, vision language models, and real-time analytics—is addressing these challenges.

The NVIDIA Blueprint for smart city AI, unveiled in June at GTC Paris, integrates Omniverse libraries, synthetic data generation, and video analytics AI agents. Updates include new Cosmos world foundation models and VLMs for photorealistic synthetic data and physical reasoning, alongside the latest Video Search and Summarization (VSS) blueprint. These tools enable rapid simulation of urban conditions and real-time vision AI, streamlining workflows for cities worldwide.

In Raleigh, North Carolina, Esri's geospatial map uses NVIDIA's video analytics AI agent to process live sensor data, allowing operators to optimize traffic flows and improve infrastructure design. The city will leverage the blueprint to automate streetlight timing, reducing road delays and carbon emissions. Milestone Systems is introducing generative AI to its XProtect platform, powered by NVIDIA Cosmos Reason VLMs trained on 75,000 hours of compliant traffic video. This could cut operator alarm fatigue by up to 30%, with cities like Dubuque and Genoa planning trials.

Linker Vision, building on success in Kaohsiung City where it slashed incident response times by 80%, is deploying the blueprint in Ho Chi Minh City and Danang. Using AVES Reality's 3D digital twins via NVIDIA Omniverse, it simulates traffic and construction to boost urban mobility. Smart Dublin collaborates with Bentley Systems' Cesium platform and Omniverse for real-time micromobility data visualization, while VivaCity uses Jetson and Metropolis platforms on AI sensors to analyze cyclist and pedestrian behavior.

Deloitte applies AI to automate street inspections, using Cosmos Predict and Transfer to generate realistic videos of intersections under various conditions, evaluated by Cosmos Reason for improvements. Hardware partners like AAEON, Advantech, and Dell Technologies showcase demos on RTX PRO Servers, DGX Spark, and Jetson Thor modules. Additional exhibits include Akila's digital twins for Monaco-Monte-Carlo train stations and K2K's real-time analytics for roadway safety.

As urban populations grow, NVIDIA's ecosystem is proving that physical AI can make cities more responsive, efficient, and sustainable. The blueprint's availability will further accelerate this transformation, with partners leading the charge in global smart city innovation.