By Alex Smith |
NVIDIA and its partners have showcased a range of artificial intelligence applications for urban development at the Smart City Expo World Congress in Barcelona, demonstrating how AI-driven digital twins, simulation and real-time data processing can be used in infrastructure and public services.
The Smart City AI Blueprint combines several of NVIDIA’s technologies – including the Omniverse platform for 3D simulation, Cosmos world foundation models for synthetic data generation, and Metropolis tools for computer vision and analytics. Together, these systems allow city planners and developers to model real-world conditions, train vision AI agents and deploy predictive tools for traffic management, energy efficiency and safety monitoring.
Recent updates to the blueprint include enhanced Cosmos Predict, Transfer and Reason capabilities. These models can generate photorealistic synthetic video data, replicate complex weather and lighting conditions, and simulate physical reasoning for autonomous systems. NVIDIA said the updates enable faster training of AI models used for video summarisation, object detection and situational awareness in live environments.
NVIDIA also highlighted the technology partners already using these systems in practical deployments during the Smart City Expo World Congress. Esri is developing an AI-powered geospatial mapping tool for Raleigh, North Carolina, that integrates camera feeds to create live maps of traffic, infrastructure issues and environmental data. The platform enables operators to respond to incidents and adjust signal timing in real time, reducing congestion and emissions.
Milestone Systems is integrating generative AI into its XProtect video management platform, using NVIDIA’s Cosmos Reason models trained on 75,000 hours of regional traffic video. The new features automate video review, produce incident summaries and filter false alarms, reducing operator workload by up to 30 per cent.
In Vietnam, Linker Vision is using the Smart City AI Blueprint to build digital twins of Ho Chi Minh City and Danang, allowing authorities to simulate traffic patterns, construction schedules and emergency response scenarios. The company reported up to 80 per cent faster incident detection in a similar deployment in Kaohsiung, Taiwan.
In Ireland, Bentley Systems and VivaCity are working with Smart Dublin to capture multimodal transport data and integrate it into a real-time digital twin. Using NVIDIA Jetson and Metropolis platforms, AI sensors track cyclists, pedestrians and vehicles to improve road safety and infrastructure planning.
Deloitte is using NVIDIA Cosmos Predict and Transfer models to automate street inspections and test pedestrian safety under different environmental conditions such as fog, rain and low light.
NVIDIA will continue to demonstrate its Smart City AI Blueprint and partner use cases at the final day of The Smart City Expo World Congress.