AI Impact on Cities

What is the New Urban Infrastructure Layer?

THE SHIFT

Cities have always been in competition with each other. First on geography and trade routes. Then on industry, labor, and scale. More recently on finance, talent, lifestyle, and global connectivity. Different places have chosen different lanes. Some win on affordability, some on institutions and innovation, some on regulation and trust, some on being a magnet for capital and people. In every era, the winners weren’t just the places with the best narrative. They were the places that could deliver the enabling conditions behind that story.

AI raises the importance of those enabling conditions because it is both energy-hungry and increasingly embedded in everyday operations.

THE PREDICTION

Training large models requires massive, concentrated power, specialist cooling, and high-capacity data connectivity, which disproportionately pulls investment into a smaller set of locations that can deliver grid headroom and connection timelines. This is concentrated among a small pool of hyperscalers; however, the use of AI is not - it affects every business. The real-world applications of AI include customer interactions, design, planning, forecasting, security, automation, and real-time decision support. Much of that workload needs high uptime, secure data flows, and lower latency. This tends to concentrate in the most connected districts and infrastructure-rich sub-regions of a city, rather than dispersing evenly.

This raises the importance of resilient digital infrastructure, edge compute, and local capacity that supports factories, logistics networks, hospitals, transport systems, offices, and city operations. Investment will tend to cluster in cities that can deliver certainty and speed: power availability, fast connections, clear permitting, available land, and credible infrastructure investment plans.

Some places will focus on hosting large-scale compute and data infrastructure. Others will focus on supporting AI-intensive users: advanced manufacturing, life sciences, logistics coordination, finance, and high-value services that depend on reliable compute and data. For many mainstream sectors, AI won’t require a dedicated data center in the city, but it will increase baseline demand for resilient connectivity, secure building systems, and higher electrical capacity in key districts.

REAL ESTATE IMPLICATIONS

That has direct implications for real estate and land strategy. Assets and zones that can support digital intensity gain strategic value: data centers and edge nodes, but also AI-enabled offices, labs, industrial automation sites, and operational control space. Locations near substations, fiber corridors, and resilient infrastructure become more investable, while planning will face sharper trade-offs around land use, energy allocation, and community impact. Cities that can coordinate utilities, planning, and investment will be able to shape where these clusters land, and how the benefits are captured.

For city governments, the opportunity is to treat power, fiber, land, and permitting as economic development tools, reducing uncertainty and making the city ‘buildable’ for the next wave of growth. Power can come from many places: grid upgrades, renewables plus storage, flexible gas, nuclear and SMRs, and over a longer horizon potentially fusion. But the near-term differentiator is execution. The cities that win in the AI economy won’t only market innovation; they will deliver capacity.

Why Will Speed of Permitting Become Competitive Advantage?

THE SHIFT

Cities have always faced change, but the built environment moves slower than the economy. In past industrial shifts, businesses could reposition faster than districts could. Buildings have long amortization horizons and resist rapid rewiring. That mismatch is exactly why permitting speed matters in the AI era: when operating models shift faster, the cost of inaction rises.

AI amplifies this in two ways. First, it accelerates structural change in how space is used (offices, retail, logistics, healthcare, education). Second, it increases economic churn: decisions, product cycles, and competitive moves update more frequently. Cities that can re-permit and re-purpose quickly will absorb that change sooner, even if physical turnover still takes time. Cities that can’t will need to carry more stranded assets, slower housing delivery, and delayed reinvention of districts.

This is where planning speed becomes competitive advantage; because speed reduces uncertainty for capital. We already have precedents. Cities like Houston, with permissive zoning and faster ‘by-right’ delivery, have historically been able to add supply quickly and keep housing more affordable than many constrained peers. Tokyo’s rules-based approach and high baseline permissions have also supported sustained housing delivery at scale. Those are not perfect analogies, but they show how speed and certainty can materially shape outcomes.

THE PREDICTION

AI will accelerate the planning process perhaps more than many expect, especially for smaller and standardized applications. Many of the inputs are document-heavy and rules-based: code checks, policy alignment, design compliance, precedent comparison, and consultation workflow. AI will automate large parts of that - and applicants will use their own AI tools to test ‘likelihood of success’ before they submit. In some categories, planning begins to disappear as a process and becomes closer to validation: ‘if it meets the plan and the code, it gets consent’. The human focus shifts upward, away from paperwork and toward politics, trade-offs, and community consent in the genuinely contested cases.

CREATING A PLAYBOOK

The practical playbook for cities is clear. First, expand ‘by-right’ permissions where proposals align with the plan; Second, standardize applications into machine-readable templates so reviews can be automated. Third, publish decision timelines and performance metrics. Fourth, build a modern planning operating model: a single digital intake, integrated utility checks, and AI-assisted triage that routes only edge cases to scarce human time. And fifth, build the political culture that makes rule-based consent legitimate: transparency, clear trade-offs, and trust that ‘by-right’ isn’t ‘by-stealth’.

What Role Will Digital Twins Play In Cities?

THE SHIFT

For years, city ‘digital twins’ were often treated as glossy replicas: good for visualization, but weak for day-to-day decisions. The binding constraint was not sensors or software, it was usability: too much fragmented data, too little standardization, limited interoperability across systems, and too few teams able to turn a model into action fast enough to create value. AI changes that by making the data legible and the workflow executable, translating noisy data into predictions, recommendations, and automated routines.

THE PREDICTION

Over the next decade, the most valuable twins won’t be 3D models. They’ll be operational platforms. They will fuse transport, utilities, asset condition, weather, and demand signals to run the city more predictably. The value sits in a set of repeatable use cases:

• First, infrastructure operations. Predictive maintenance for roads, bridges, water networks, and substations - prioritizing interventions before failure and reducing emergency call-outs.
• Second, energy and grid performance. District-level demand forecasting, demand response, and smarter load management that reduces peak stress and improves resilience.
• Third, mobility. Traffic optimization, incident response, and better scheduling of street works to minimize disruption.
• Fourth, climate risk. Heat and flood modeling that informs zoning, retrofit priorities, insurance pricing, and emergency planning.
• Fifth, permitting and development. Faster validation of proposals against constraints (utilities capacity, transport impact, carbon targets, and resilience rules) - reducing uncertainty for capital.
• And sixth, public safety and service delivery: faster routing of crews, better allocation of scarce resources, and earlier detection of emerging issues.

REAL ESTATE IMPLICATIONS

The real estate implications are direct. Resilience becomes more measurable: flood risk, heat stress, outage exposure, and mobility reliability are factors of underwriting. Performance becomes more optimizable - buildings that can share data and respond (HVAC control, energy flexibility, occupancy management) can participate in district systems and reduce operating cost. That creates a premium for assets with strong data capture and controllable building systems; and it increases obsolescence risk for stock that can’t integrate.

This becomes a competitive strategy for cities because it increases delivery certainty. A city that can quantify constraints up front can streamline approvals, coordinate utilities, and target capital spending more effectively.

The private sector taps in through clear data and integration standards. Developers can design to quantified constraints (utilities capacity, resilience thresholds, mobility impacts), reducing redesign and delay. Landlords can connect buildings into district systems, enabling demand response and lowering operating costs. Insurers and lenders can price risk using better, more current inputs, improving underwriting and reducing uncertainty. Operators can use shared signals to cut downtime and coordinate maintenance.

Whether a city’s twin is ‘open’ will vary. Some cities will publish selected data layers and APIs to encourage third-party tools, innovation, and investment. Others will keep parts restricted, especially around security, critical infrastructure, and personally identifiable data, while still enabling controlled access for verified partners.

Either way, the winners will be the cities that treat the twin as shared infrastructure, with clear rules on access, privacy, and commercial use, so that public and private systems can actually work together.

How Will the City’s Activity Maps Get Redrawn?

THE SHIFT

Cities are not just defined by buildings. They are defined by patterns of presence and interaction: where people go, how often, and for what. For a century, the dominant pattern has been predictable: concentrate work centrally, cluster higher-order services in a few districts, and disperse housing outward.

AI changes the mix of trips within the city by reducing the penalty of distance for some activities and raising the value of being together for others. As work becomes more digitally supported (planning, drafting, analysis, coordination) fewer interactions require a specific place at a specific time. That makes presence more intentional. City centers don’t lose relevance; they concentrate further around what density is uniquely good at: complex collaboration, specialist services, culture, and experiences that work better with footfall. The change is gradual: routine, low-stakes trips reduce first, and activity spreads across more centers within the same city.

THE PREDICTION

Residential neighborhoods shift from being primarily ‘where you sleep’ to being more functional in the daytime. When coordination becomes cheaper, services that used to need scale can run in smaller, local formats. You see more micro-nodes: pickup and return points, repair and recommerce drop-offs, screening and pop-up clinics, tutoring and skills sessions, and on-demand local mobility and care support that can be routed intelligently. Streets and third spaces matter more because they carry more of daily life: errands, health, learning, and social contact happening closer to home, more often, and in shorter bursts.

The consumer city shifts too. AI makes discovery and selection easier, but it also makes fulfillment and aftersales faster, which changes what physical places are for. Shopping leans toward experience, immediacy, service, and entertainment, while the ‘invisible’ layer grows - returns, repair, and last-mile functions that sit behind the scenes but shape where space is needed. Leisure becomes easier to find and easier to organize: better recommendations, smoother booking, and more dynamic programming that can lift footfall in the right places.

As coordination and programming get cheaper, cities can activate more public space more often, turning civic life from occasional events into a reliable, bookable layer of everyday services and social activity. This is where a new type of civic center emerges; not just a building, but a stitched-together network of indoor and outdoor places. AI makes public space more adaptive: libraries that double as skills studios and maker spaces, community hubs with bookable rooms and hybrid capability, and ‘care and advice’ points linked to city services. Outside, streets and squares become more usable, managed for markets, performances, sport, pop-up learning, and seasonal uses, with better booking, wayfinding, lighting, and maintenance. Civic space becomes a platform for participation, not just a destination.

REAL ESTATE IMPLICATIONS

The real estate implication is reallocation inside the city. Secondary centers strengthen: town centers, rail nodes, and mixed-use hubs that combine flexible space with services. Ground floors evolve toward service, care, and logistics-supporting uses (pickup and return, repair, community health, learning), not just traditional retail. Districts that can flex between work, learning, health, and leisure become more valuable than single-use monocultures.

For investors, the opportunity is to back the nodes that capture these new patterns: amenity-rich centers, transit-connected districts, and assets that can be re-tenanted and reprogrammed as demand shifts. For city government, the opportunity is to make reallocation investable: zoning that enables mixed-use intensity, simpler conversion pathways, and public realm investment that supports civic activity indoors and out. The cities that win won’t fight change, they’ll design for new patterns of presence.

How Will Mobility and Access Change?

THE SHIFT

Urban mobility has always been constrained by two things: limited network capacity and imperfect coordination. AI improves the coordination layer. Over the next decade, routing, pricing, scheduling, and real-time control become more dynamic across public transport, roads, freight, and micromobility. The immediate gain is reliability: fewer wasted minutes, smoother interchanges, and better use of existing infrastructure.

THE PREDICTION

That optimization will change modes and nodes. You’ll see more ‘managed curb’ space for pickup, deliveries, and short-stay access. You’ll see more micromobility and demand-responsive shuttles feeding rail hubs. You’ll see freight and service traffic routed more intelligently to avoid peaks. Autonomous vehicles will arrive first in controlled environments (campuses, ports, logistics corridors, airports) and then expand as regulation and safety mature. Electrification accelerates, and charging becomes critical spatial infrastructure: depot charging for fleets, rapid charging on main corridors, and slower ‘top-up’ charging embedded in car parks, supermarkets, and mixed-use hubs.

These shifts reshape legacy interchanges. Stations and hubs become higher-throughput transfer points, not just ticket halls, because AI reduces the friction of switching modes. Better prediction and live coordination enables timed connections, dynamic platform management, and clearer wayfinding. Managed curb space reduces the chaos of pickup and drop-off. If cities run a mobility ‘digital twin’, a live model of flows, incidents, and constraints, they can pre-empt bottlenecks, coordinate streetworks, and reroute demand in real time rather than reacting after congestion has formed.

This improves congestion and pollution in practical ways. Congestion falls when you cut ‘wasted’ movement and improve flow: fewer drivers circling for parking, fewer stop-starts at signals, and fewer empty miles from poorly routed fleets. Autonomous vehicles can help where they reduce shockwaves (tighter spacing, fewer harsh accelerations, and more consistent speeds). But the bigger win is system coordination. Pollution falls when traffic flows more smoothly and when electrified fleets shift charging to off-peak patterns, reducing idling and local emissions in busy corridors.

REAL ESTATE IMPLICATIONS

The second and third-order real estate impacts are as follows. As travel becomes more predictable, the practical size of a city expands: some suburbs become newly ‘commutable’ because reliability improves. That can reprice residential demand and accelerate development around rail nodes and high-quality interchanges. Mobility corridors attract more mixed-use clustering (workspace, services, education, health, and hospitality) because access becomes more guaranteed. And car parks and road-facing assets evolve: less long-stay commuter parking, more charging, servicing, short-stay turnover, and pickup space as the arrival experience becomes part of the asset.

Commuting patterns will keep evolving as work becomes more distributed. But even in that world, mobility matters more, not less, because cities compete on how easily people can reach the places they still choose to go. For city government, the opportunity is to treat mobility as managed infrastructure: build a live ‘system view’ of flows, digitize and price curb space, coordinate traffic signals and street works in real time, and accelerate EV charging and shared last-mile capacity so reliability improves without endlessly adding road capacity.