The urban population is expected to double by 2050. That means that by the middle of the century, about 2.5 billion people will be added to urban areas, promoting the need for more sustainable urban planning and public services. Cities around the world are turning their eyes to digital twins and AI agents for city planning scenario analysis and data-driven operational decisions.
But building a digital twin for the city and testing smart city AI agents within it is a complex, resource intensive effort full of technical and operational challenges.
To address these challenges, Nvidia today announced Smart City AI's Nvidia Omniverse Blueprint. It is a reference framework that combines the Nvidia Omniverse, Cosmos, Nemo and Metropolis platforms to bring the benefits of physical AI to the entire city and its critical infrastructure.
Using Blueprints, developers can build simulation-enabled or city optical aeristic digital twins to build and test AI agents that help monitor and optimize city operations.
Large companies like XXII, Aves Reality, Akila, Blyncsy, Bentley, Cesium, K2K, Linker Vision, Milestone Systems, Nebius, SNCF Gale & Connection, Trimble and Younite AI were the first to use the new blueprint.
https://www.youtube.com/watch?v=9k3y6krgpxo
Smart City AI's Nvidia Omniverse Blueprint
Smart City AI's Nvidia Omniverse Blueprint provides the complete software stack needed to accelerate the development and testing of physically accurate digital twin AI agents in cities. It includes:
The Blueprint Workflow consists of three important steps: First, developers will use aviation, satellite, or map data using Omniverse and Cosmos to create Simready Digital Twin locations and facilities. Second, AI models such as computer vision models and VLMS can be trained and fine-tuned to improve the accuracy of vision AI use cases using Nvidia Tao and Nemo curators. Finally, real-time AI agents with these customized models are deployed to alert, summarise and query camera and sensor data using Metropolis VSS Blueprints.
The Nvidia Partners' ecosystem powers smart cities around the world
The Smart City AI blueprint allows a large ecosystem of partners to use a single workflow to build and activate digital twins for smart city use cases, leveraging Nvidia's technology and unique combinations.
SNCF Gares & Connexions, which operates a network of 3,000 railway stations across France and Monaco, has deployed digital twin and AI agents to enable real-time operational monitoring, emergency response simulation and infrastructure upgrade planning.
This allows each station to analyze operational data such as energy and water use, allowing predictive maintenance capabilities, automated reporting, and GDPR-compliant video analysis for incident detection and crowd management.
With the deployment of physical AI at Mona Commone Carlo and Marseille Stations of ecosystem partners Akira and XXII's Omniverse, metropolitan, solutions, and physical AI at SNCF Galle & Connection's Physical AI at Monacom Monte Carlo and Marseille Stations, SNCF Galle & Connection reduces the completion rate of preventive maintenance by 100%.
The city of Palermo in Sicily is using partner K2K's AI agents and digital twins to improve public health and safety by helping city operators process and analyze footage from more than 1,000 public video streams at a rate of nearly one second.
Sicily tapped AI agents for K2K are built with VSS's NVIDIA AI Blueprint and Nebius's cloud solution, but can interpret and act on video data to provide real-time alerts on public events.
To accurately predict and resolve traffic accidents, K2K uses the COSMOS World Foundation model to generate synthetic data to simulate various driving conditions. K2K then uses the data to fine-tune the VLMS with the NEMO curator in its AI agent. These simulations allow K2K's AI agents to produce over 100,000 predictions per second.
The Milestone System has launched Project Hafnia in collaboration with NVIDIA and European cities. This is an initiative to build an anonymized, ethically sourced video data platform to develop and train AI models and applications while maintaining regulatory compliance.
Using the combination of Cosmos and Nemo curators on the Nvidia DGX Cloud and Nebius' sovereign European cloud infrastructure, Project Hafnia enables European-compliant training and fine-tuning of video-centric AI models, including VLM, for a variety of smart city use cases.
The first deployment of the project, held in Genoa, Italy, features one of the world's first VLM models for intelligent transport systems.
Linker Vision was one of the first to partner with NVIDIA to deploy Smart City Digital Twins and AI agents in Kaohsiung City, Taiwan, powered by Omniverse, Cosmos and Metropolis. Linker Vision has teamed up with digital twin company Aves Reality to bring aerial images of cities and infrastructure into 3D geometry and ultimately to the Simready Omniverse Digital Twins.
Linker Vision's AI-powered applications built, trained and tested visual AI agents on digital twins before being deployed in physical cities. Currently, it is scaled to analyze 50,000 video streams in real time using generated AI to understand and narrate complex urban events such as floods and traffic accidents. Linker Vision provides timely insights to dozens of urban sectors through a single, integrated AI-powered platform, breaking silos and reducing incident response times by up to 80%.
Bentley Systems is taking part in an effort to bring physical AI into the city using Nvidia Blueprint. Cesium, an open 3D geospatial platform, provides the foundation for visualizing, analyzing and managing infrastructure projects, porting the Digital Twins to the omniverse. The company's AI platform Blyncsy uses synthetic data generation and metropolitan areas to analyze road conditions and improve maintenance.
Trimble, a global technology company that enables critical industries such as construction, geospatial and transportation, is exploring ways to integrate Omniverse Blueprint components into real-life capture workflows, integrating the Trimble Connect Digital Twin platform for mapping survey applications for smart cities.
Younite AI, developer of AI and 3D Digital Twin Solutions, employs blueprints to accelerate the development pipeline, quickly moving from operational digital twins to large-scale urban simulations, improving synthetic data generation, integrating real-time IoT sensor data, and enabling AI agents to be deployed.
Learn more about Smart City AI's Nvidia Omniverse Blueprint by participating in this GTC Paris session or watching on-demand videos after the event. Please sign up to be notified when a blueprint is available.
Please take a look Nvidia GTC Paris Keynote From Jensen Huang, founder and CEO of Vivatech's Nvidia GTC Paris Session.
