How urban planners in Barcelona and others are using supercomputers and AI to create digital twins of their cities to address issues like traffic congestion
Joy Lo Dico / Financial Times :
Context & Ripple Effects
Barcelona’s work sits within a broader municipal shift toward AI-enabled public services, alongside city efforts to establish rules for AI in public services. The distinctive element here is applying supercomputing and AI to a city-scale operational model rather than a single service.
The same design-and-planning ecosystem is adopting AI in adjacent workflows, from generative tools in architectural design to AI-assisted environmental optimization. Digital twins extend that logic from creating individual spaces to testing interventions across an urban system.
First-order effects
- Urban planners in Barcelona and peer cities gain a simulation layer for evaluating traffic-congestion interventions before applying them in the physical city.
- The projects make access to supercomputing, AI systems and the underlying city data a more central part of planning operations.
Second-order effects
- Municipal AI governance becomes more consequential: city rules must cover not only public-facing services but also models that can inform transport and planning decisions.
- Infrastructure and planning teams face stronger incentives to connect transport data and other operational inputs to shared models; comparable AI uses for fuel-efficient driving and grid monitoring show the adjacent operational opportunity.
Third-order effects
- If these systems become embedded in planning, urban policy could shift toward continuously modeled, data-dependent decision processes rather than periodic, static planning exercises.
- That shift would make public-sector compute capacity, data stewardship and model accountability durable constraints on which cities can deploy such tools effectively.
The trend: City governments are moving AI from discrete public-service applications into digital representations of physical infrastructure for ongoing operational decision-making.