
Tech • AI • Robotics • Game
NVIDIA outlined how cities can move from digital ambition to implementation by using AI for planning, operations and citizen services, with examples from France, Italy and Spain.
Cities have long been shaped by transport, water, roads and power, but the next decade is expected to center on building intelligence infrastructure. The goal is to turn urban data into better services for residents while also creating an economic flywheel that attracts talent, startups and investment.
Municipalities now generate vast streams of information from utilities, pollution monitors, traffic systems, EV charging, mobile devices, drones and satellites. The central challenge is that the gap between available data and data actually used for decision-making remains wide, prompting cities to invest in systems that can contextualize and act on that information.
The most advanced deployments fall into three categories: digital twins for planning, vision AI for operational management, and agentic AI for citizen services. These uses are spreading across Europe, the Middle East and Africa as cities test what is technically and economically feasible at scale.
In the Dijon area, the Department of Côte-d'Or has worked with local technology developers to create enriched digital models for more than 500 buildings. These are not simple 3D replicas: they include maintenance histories, utility records and other operational data, allowing authorities to improve current building management and simulate future planning scenarios more cheaply and quickly than before.
Multi-dynamics simulation, long associated with sectors such as automotive, healthcare and robotics, is now being applied to cities. In France, companies including Aquila and Think Deep are using NVIDIA-based models on sovereign cloud infrastructure, reflecting growing interest in keeping sensitive public data under tighter regional control.
Palermo has developed what amounts to a city brain by combining sensor feeds and other urban data sources in a central operations capability. The system is designed to cut emergency response times, improve traffic flow, strengthen safety functions and support areas such as waste management in a dense historic city where street layouts complicate mobility.
Palermo’s approach paired system integrators, infrastructure providers such as Dell, and local software specialists including K2K. That model reflects a broader pattern in urban AI: successful rollouts depend not just on hardware, but on local implementation partners and software libraries that can accelerate development.
In Barcelona, the company 1MillionBot has built an agentic platform using Alia, a language model developed on local supercomputing infrastructure in collaboration with government. The model is tuned for Spanish and regional linguistic variants including Galician and Basque, enabling digital assistants to book appointments, guide residents through procedures and handle service inquiries in more accessible language.
Public-sector AI projects are placing heavy emphasis on sovereignty, permissions and secure access because they often involve sensitive citizen information. Current AI architectures are increasingly designed to let cities retain control over data, applications and user rights while still deploying advanced agents and automation.
The main bottleneck is often not funding but the availability of technical capabilities. Cities are increasingly creating centers of excellence and appointing chief AI roles to bring together cloud providers, telecoms, infrastructure operators and solution developers around shared use cases.
The argument presented to city leaders is that where compute is built, talent and startups tend to follow. With many new companies now born as AI-native businesses, urban areas that secure power, infrastructure, models and applications across the full AI stack may gain disproportionate benefits in innovation, job creation and long-term competitiveness.
Large AI investment figures can obscure the fact that deployments can begin with relatively compact systems and a single high-impact use case. The emphasis is on modular implementation, quick returns on investment and choosing projects that visibly improve community outcomes before scaling across departments.
The emerging urban AI model is less about futuristic branding than about building usable systems for planning, operations and resident services. Cities that combine data control, technical skills and local compute capacity are likely to gain both better public services and a stronger innovation economy.
Ask a question