Why Smaller Cities Are Becoming Europe’s New Data Centre Hub

A proposed data centre can arrive at a smaller city hall looking like the investment opportunity of a generation. The project may carry a multibillion-euro price tag, promise a place in Europe’s growing artificial intelligence economy and turn an overlooked industrial site into infrastructure of international importance. Yet the investment figure says surprisingly little about what the development will contribute to the city around it.

Europe’s data-centre map is beginning to move beyond Frankfurt, London, Amsterdam, Paris and Dublin. These established markets remain important, but access to land and electricity has become increasingly constrained. Large facilities can require hundreds of megawatts of connection capacity, while suitable sites must also offer resilient fibre networks, reliable permitting and enough physical space for future expansion.

That combination is becoming difficult to secure in major metropolitan areas. Developers are therefore looking towards smaller cities, industrial regions and metropolitan hinterlands where land is more readily available and grid connections may be achievable within a commercially realistic timeframe.

For local governments, this creates an unusual opportunity. Smaller cities that spent years competing for factories, logistics centres and corporate offices are now being considered for some of Europe’s most valuable digital infrastructure. The municipalities that benefit, however, will be those that understand what they are being asked to host—and negotiate accordingly.

AI Is Changing What Developers Need From a Location

Traditional data centres were often built close to major business centres and internet exchanges because speed and proximity to users were central considerations. Artificial intelligence is changing part of that calculation.

Training large AI models involves vast computational workloads, but much of this processing does not have to take place in the centre of Frankfurt or another established digital hub. Where latency is less critical, developers have greater freedom to place computing capacity near available electricity, renewable generation and large development sites.

The result is a new class of project that resembles industrial infrastructure as much as commercial property. In Germany, plans reported in 2026 include a facility of up to 200 megawatts in Hamm, another 200-megawatt development in Lübbenau and a project in Rackwitz that could eventually reach 500 megawatts. These are not server rooms attached to office buildings. Their energy requirements can be comparable to those of major industrial operations.

Demand is likely to continue. Bitkom expects the share of German data-centre capacity dedicated to AI to rise from approximately 15 percent to 40 percent by 2030. Cloud migration is progressing at the same time, leaving operators searching for sites that can accommodate both greater capacity and higher computing density.

Smaller cities may offer the land, but land alone is rarely the decisive advantage. The scarce asset is increasingly electricity: not merely access to a power line, but a confirmed connection capable of supporting the project within an acceptable period. In Frankfurt, the waiting time for a large connection can extend beyond a decade. A well-located regional city with available grid capacity can therefore become more attractive than an internationally recognised data-centre market.

A Large Investment Does Not Guarantee a Large Local Economy

The scale of the construction budget can dominate the political discussion. A project worth several billion euros is easy to present as an economic breakthrough, particularly in a region seeking new investment after the decline of traditional industries.

The local outcome is more complicated. Data centres create substantial activity during construction and require specialist personnel once operational, but they do not generate permanent employment on the scale of an automotive plant or a large manufacturing campus. Many of the highest-value contracts may also go to companies based elsewhere.

This does not make a data centre a poor investment for the host city. It does mean that municipal leaders should assess the project through a different lens. The relevant questions concern tax revenue, infrastructure obligations, land use, energy supply, water demand and the services that can develop around the facility.

A city that approves a data centre without a broader economic plan may gain an impressive building and relatively little integration with the local economy. A more ambitious municipality can use the investment to attract cloud providers, cybersecurity firms, technical maintenance companies, research partnerships and training programmes. Reliable digital infrastructure may also strengthen the city’s case when competing for other technology-intensive businesses.

None of these benefits will appear automatically. They depend on decisions made during site selection, planning and negotiation, often years before the first servers are installed.

Electricity Must Be Treated as a Strategic Urban Resource

The arrival of a major data centre changes more than one development site. It can affect the city’s entire energy strategy.

The same grid capacity may also be needed for new housing, public transport electrification, heat pumps, electric-vehicle charging and the decarbonisation of local industry. Allocating hundreds of megawatts to one facility is consequently not a routine technical decision. It determines which other developments may be connected, how quickly and at what cost.

Municipal planning departments and grid operators have not always worked closely enough to manage this competition. The expansion of AI infrastructure makes that separation increasingly untenable. Cities considering data centres need a shared map of available capacity, planned grid reinforcement and future demand from other sectors.

Developers can contribute to a more resilient system through batteries, on-site renewable generation and flexible power arrangements. Data centres may be able to reduce demand during periods of grid stress or use backup systems to provide balancing services. Such claims require detailed examination, however, because theoretical flexibility is not the same as a binding operational commitment.

The city should know how much electricity the site will require at each stage of its development, which grid upgrades will be necessary, who will finance them and what happens if the facility expands. Without that information, the municipality may approve a project whose wider infrastructure consequences remain unclear.

Waste Heat Can Be Valuable, but Only Where Someone Can Use It

Data-centre developers increasingly present waste-heat recovery as part of their sustainability case. The principle is attractive: servers generate heat, which can be redirected into homes, public buildings or industrial processes instead of being released into the environment.

In practice, the value of that heat depends on location. A facility situated far from a district-heating network cannot become an urban heat source simply because the technology exists. The temperature of the recovered heat, the distance to potential clients, demand throughout the year and the cost of heat pumps and pipelines all influence whether the system is commercially viable.

Heat reuse therefore needs to be planned before the data centre is built. The municipality, energy utility and operator should identify potential users, calculate the necessary investment and agree who will carry the financial and operational risk. Reserving a possible connection for an unspecified future network is not equivalent to delivering usable heat.

This is where smaller cities can hold an advantage. Former industrial sites may already be close to energy infrastructure, commercial heat users or neighbourhoods suitable for district heating. A city can also coordinate the data-centre timetable with housing developments, municipal buildings or the conversion of an existing heat network.

The project then becomes part of an urban energy system rather than an isolated consumer attached to the grid.

Water and Land Require the Same Scrutiny

Electricity receives most of the attention, but cooling can also create significant water demand. The amount varies widely according to climate, technology and operating model, which makes broad assurances of efficiency inadequate. Cities need project-specific estimates showing expected consumption under normal and extreme weather conditions.

This is particularly important in regions already experiencing pressure on water supplies. A technology company’s sustainability targets do not remove the municipality’s responsibility to examine how the facility will interact with households, agriculture and local industry during drought or heatwaves.

Land use presents another trade-off. Data centres require large, secure sites and may deliver limited street-level activity. Placed carelessly, they can occupy valuable urban land that might have supported housing, mixed-use development or employment with a greater local workforce.

Industrial zones, brownfield sites and areas with existing electricity infrastructure are generally more suitable than locations that interrupt residential neighbourhoods or consume scarce central land. Architecture and landscaping still matter: a large windowless facility, surrounded by fencing and generators, can shape the surrounding district for decades.

Cities Need Their Own Data-Centre Strategy

The balance of power may appear to favour the developer, especially when several regions are competing for the same investment. Yet operators also need political support, predictable planning and long-term certainty. A city that has already defined its expectations may prove more attractive than one that promises speed but lacks a coordinated position.

A municipal data-centre strategy should establish where facilities are appropriate, how projects will be assessed and which local contributions are expected. It should bring together planning, economic development, energy, water, environmental protection and emergency services rather than leaving the decision to a single department.

The strategy also needs to distinguish between different types of facility. A regional cloud centre, an AI training campus and an edge data centre serve different functions and place different demands on the city. Treating them as one category prevents a meaningful assessment of their costs and benefits.

Public communication should begin before a project reaches its final approval stage. Residents are more likely to distrust a development when its benefits are described in abstract language while its electricity use, water requirements and physical scale remain difficult to establish. Transparent figures and enforceable commitments provide a stronger case than promises of joining the digital future.

Frankfurt and Europe’s other major data-centre markets will not lose their relevance. Their connectivity, technical expertise and established ecosystems cannot simply be recreated elsewhere. The next phase of growth, however, will be more geographically dispersed, driven by the search for land, energy and faster development.

That shift gives smaller cities a position they have rarely held in the technology economy. They are no longer only competing to attract an investment; they can help determine what responsible digital infrastructure should look like. The opportunity is considerable, but the best outcome will not belong to the city that offers the cheapest land or the quickest approval. It will belong to the city that knows which conditions to set before the servers arrive.