How Town Planning Works | TPW-0076 — The Thermal Network: How District Energy, Waste Heat and Underground Pipes Turn Neighbouring Buildings Into One Energy System

A city can throw away enough heat to warm another part of itself.

Data centres reject heat while servers run. Factories release heat from industrial processes. Wastewater carries thermal energy beneath the street. Incineration, power generation, refrigeration and large cooling systems move heat from one place to another. At the same time, nearby homes, hospitals, hotels and offices pay to create heat—or, in hot climates, pay to remove it.

Normally, each building solves its own thermal problem.

A district-energy network asks a different question: what if buildings were planned as parts of one thermal system?

District heating and cooling networks use shared plants, pipes, heat exchangers, storage and controls to move thermal energy among multiple buildings. They can use conventional energy sources, renewable electricity, geothermal energy, recovered industrial heat, wastewater heat, data-centre heat or combinations of several sources. The technology is well established. The planning challenge is spatial.

Heat is valuable only if the source, the users and the network can be connected at the right scale, temperature, time and cost.

That planning question is becoming more important. The American Planning Association’s March 2026 PAS QuickNotes on data-centre waste-heat recovery gives planners explicit guidance for treating rejected heat as a potential community resource rather than an unavoidable by-product. Dublin City Council’s current district-heating project plans to recover up to 90 megawatts of waste heat from the Dublin Waste to Energy facility and move it through underground pipes to dense Docklands and Poolbeg development. In Singapore, the Urban Redevelopment Authority’s 2026 long-term planning work explicitly highlights district cooling and coordinated infrastructure as part of systems-level urban planning, while current development-control incentives encourage building owners to connect to district or centralised cooling networks. Across Europe, the 2026 District Heating and Cooling Market Outlook reports continuing growth in networks that integrate renewable energy, waste heat, electrification and thermal storage.

The reader job of this article is therefore distinct from general energy planning: how does a town decide where a shared thermal network is actually worth building, what it should connect, and how to keep the network useful as the district changes?

Heat has geography

Electricity can travel long distances relatively efficiently through wires. Heat is more local.

Hot or chilled water loses energy as it moves. Pipes cost money. Pumps use electricity. Streets must be opened. Connections must be installed building by building.

This means thermal networks depend on spatial relationships.

A large source of waste heat ten kilometres from sparse development may be less useful than a smaller source beside a dense hospital, housing and office district. A district-cooling plant can be efficient where many large buildings need cooling within a compact area. The same system can be uneconomic where loads are scattered.

The first planning map is therefore not a zoning map. It is a thermal-density map.

Load density is the basic economic geometry

District energy works best when substantial thermal demand exists along relatively short pipe routes.

Dense development helps because each metre of pipe can serve more floor area and more customers.

This is why Dublin City Council identifies high population density and compact development as ideal conditions for district heating, and why district cooling has become particularly useful in dense business districts and large mixed-use developments.

But density alone is not enough.

A district of efficient residential buildings may have lower thermal demand than a hospital-and-hotel cluster. A data centre may create enormous year-round rejected heat while neighbouring homes need the most heat only in winter. A tropical office district may have strong cooling demand in business hours and lower demand overnight.

The planner therefore needs a load profile, not only a floor-area count.

Anchor loads make networks possible

Many district-energy systems begin with anchor customers.

Hospitals, universities, airports, hotels, civic complexes, data centres, shopping districts and large housing estates can provide stable demand that justifies the first section of network.

Once the backbone exists, additional customers can connect.

The planning implication is important: the best first route is not necessarily the route that ultimately serves the largest area. It may be the route that connects two or three reliable loads strongly enough to make later expansion credible.

This is the thermal version of infrastructure sequencing. Build a viable spine before drawing an ambitious web.

The source map should include heat people normally ignore

A conventional energy map shows power stations, substations, fuel infrastructure and renewable resources.

A thermal-network map should add rejected heat.

data centres;
waste-to-energy facilities;
industrial plants;
large refrigeration systems;
wastewater treatment plants;
sewers and treated effluent;
metro tunnels and underground infrastructure;
large cooling plants;
power generation;
geothermal sources;
solar thermal installations;
surface water and groundwater where heat-pump systems are appropriate.

The source map turns a discarded by-product into a planning input.

Data centres make waste heat visible because their growth is visible

Data centres are especially interesting because they consume substantial electricity and reject heat continuously.

The American Planning Association’s 2026 waste-heat guidance notes that data centres can use several times the energy of similarly sized office buildings and that their servers require constant cooling. That thermal output can become useful when nearby buildings can accept it.

This does not mean every data centre should be connected to housing.

Waste-heat temperature may be low. Heat pumps may be required to raise it. Customers must be close enough. The data centre must provide a reliable heat stream. Contracts must address what happens if computing load changes or the facility closes.

The data centre is therefore a source candidate, not an automatic district-energy plant.

TPW-0053 — The Data Centre District remains the broader owner for data-centre land, power, water, noise and local-value impacts. The Thermal Network owns only the question of how rejected heat can become shared urban infrastructure.

Temperature quality matters as much as quantity

One megawatt of heat is not always equally useful.

Industrial processes can release high-temperature heat. Data centres often reject lower-temperature heat. Wastewater may contain a large amount of thermal energy at modest temperature.

The network needs to match source temperature with customer requirements.

Older buildings designed around high-temperature radiators may require hotter water than modern low-temperature buildings. Domestic hot water introduces another requirement. Cooling networks operate through a different temperature regime altogether.

Heat pumps can bridge temperature gaps, but they consume electricity and add capital cost.

The planner should therefore avoid mapping only energy quantity. Thermal quality belongs on the map.

Fourth- and fifth-generation networks change what can connect

Traditional district-heating systems often circulate very hot water or steam.

Newer networks increasingly use lower temperatures, improving efficiency and making low-grade heat sources easier to integrate. Some fifth-generation systems use ambient-temperature loops with building-level heat pumps, allowing buildings to exchange heat rather than simply receive it from one central plant.

This changes urban possibilities.

An office building rejecting heat during the day can potentially contribute to a shared loop while residential buildings draw heat. A supermarket refrigeration system can become a source. Cooling and heating can become two sides of the same network.

The design becomes more complex, but the planning logic becomes more interesting: neighbouring buildings are no longer isolated energy consumers. They can become thermal partners.

Mixed use can improve the load curve

District energy benefits when different buildings peak at different times.

Offices may peak during weekday business hours. Hotels and homes have strong evening and morning demand. Hospitals operate continuously. Retail has another rhythm.

A mixed-use district can therefore use shared capacity more efficiently than a mono-use district with one synchronized peak.

This gives mixed-use planning an energy-system dimension beyond street activity.

Mixed Use remains the owner for how different activities keep a place alive over time. The thermal-network question is narrower: can different thermal schedules share infrastructure more efficiently?

District cooling is the tropical version of the same planning logic

In hot climates, cooling can dominate building energy demand.

District cooling centralizes chilled-water production rather than giving every building a separate large chiller system.

Central plants can operate more efficiently, use diversity in building loads, reduce duplicated equipment and free space in individual buildings.

Singapore’s current planning framework explicitly supports this approach. URA’s systems-level planning for future districts describes centralized district cooling alongside common services tunnels, and URA’s development-control incentives allow bonus utility floor area for host developments that provide plant space serving a wider district network.

The important point is institutional: district cooling requires buildings to be planned as participants in a shared system before every parcel has independently optimized its own plant room.

The pipe corridor is the real estate problem

Thermal networks need pipes.

Pipes need space beneath streets already occupied by sewers, water, electricity, telecoms, drainage and transit infrastructure.

Installing a network after a district is complete can require repeated road opening, traffic disruption and complex utility coordination.

Installing sleeves, corridors or pipe routes during initial development can be much cheaper.

The planner therefore needs to protect the thermal route early.

The Hidden Town remains the canonical utility-corridor owner. The Thermal Network adds one specific requirement: thermal pipes need continuous routes between sources, plants, storage and customer buildings.

Street reconstruction can be the moment to build the thermal future

Opening a street is expensive.

If a road is already being rebuilt for drainage, transit or utility renewal, that project creates an opportunity to install thermal pipes or at least reserve space for them.

Capital planning should therefore compare infrastructure programmes across departments.

A district-energy project that looks uneconomic as a standalone excavation can become more attractive when coordinated with work that already requires the street to be opened.

This is a powerful example of systems-level planning: the value comes from timing projects together, not merely designing each project well.

New districts are easier than old districts

Greenfield or major regeneration areas allow pipes, plants and customer interfaces to be planned before buildings are complete.

Existing districts are harder.

Buildings may have incompatible internal systems. Owners may be reluctant to replace functioning boilers or chillers. Streets are crowded. Connection requires negotiation property by property.

Retrofit networks therefore need a transition strategy.

Start with public buildings or willing anchor customers. Coordinate connections with equipment replacement cycles. Expand when existing plant reaches end of life. Avoid forcing a building to discard a new efficient system merely to satisfy network ambition.

The network should grow around real replacement opportunities.

Connection timing can determine viability

A thermal network has high upfront infrastructure cost.

If customers connect slowly, the operator can carry debt while pipes remain underused.

This creates a sequencing problem between development and energy infrastructure.

A new district may be planned for ten thousand homes and take fifteen years to build. The district-energy plant cannot necessarily be sized for final demand on day one without creating expensive unused capacity.

Modular plants, temporary energy centres and staged pipe expansion can match infrastructure to growth.

The Time Layer owns the broader sequencing logic. The thermal network applies it to source capacity, pipe construction and customer connection.

Thermal storage separates production time from demand time

Heat or chilled water does not always need to be produced at the instant it is used.

Thermal storage can shift energy across hours or, in some systems, longer periods.

A district-cooling system can produce chilled water during periods of lower electricity demand and use stored cooling during peaks. A district-heating network can store heat from industrial or renewable sources when available.

Storage changes the planning footprint.

Tanks, pits, aquifers or other storage systems need land, underground space and appropriate geology. Large-scale thermal storage was a major focus of the 2026 Linz District Heating Forum because it can make variable renewable and waste-heat systems more flexible.

The town should therefore map storage potential alongside sources and loads.

The network can help the electricity grid

Electrified heating and cooling increases electricity demand.

If every building uses a heat pump or chiller at the same peak moment, the power grid must support that combined load.

District systems can provide flexibility through thermal storage, diverse sources and coordinated operation.

Copenhagen’s 2026 thermal-flexibility work illustrates this wider direction: buildings can become active participants whose heating demand responds to network conditions rather than passive consumers.

The future thermal network is therefore not merely a pipe system. It can become a buffer between buildings and the electricity system.

Waste heat should not justify wasteful primary activity

Recovering heat is useful.

It should not become an excuse to ignore the efficiency of the source that created the heat.

A data centre should still improve computing and cooling efficiency. An industrial plant should still reduce unnecessary energy use. A waste-to-energy facility should not be fed material that could have been avoided, reused or recycled simply because the heat has value.

The hierarchy is important: reduce avoidable energy demand first, then recover useful heat from the remaining process.

This connects with The Circular Town, which owns the broader resource-loop logic.

The carbon test needs the real energy mix

District energy is not automatically low carbon.

A network supplied by fossil fuel can be carbon intensive. Pumping consumes electricity. Heat losses matter. Backup boilers may operate during peaks.

The correct comparison is between the network’s actual source mix and the realistic alternative building-by-building systems.

A network can also improve over time. Fossil heat can be replaced by large heat pumps, geothermal energy, renewable electricity or recovered heat while keeping the same pipe infrastructure.

This source flexibility can be one of the strongest long-term advantages of district systems.

A pipe can outlive its first energy source

Energy technologies change faster than urban streets.

A well-planned network can therefore separate distribution infrastructure from generation technology.

The source in 2030 does not need to be the source in 2050.

A waste-to-energy plant may initially supply heat. Later the network can add industrial waste heat, large heat pumps, geothermal energy or seasonal storage.

The planner should protect this adaptability through plant sites, pipe capacity, interconnection points and space for future sources.

Redundancy matters because thermal networks can create dependency

Connecting many buildings to one network can improve efficiency and create a shared point of failure.

The network therefore needs resilience.

Multiple sources, looped pipes, backup plant, emergency power, sectional isolation and repair capability can reduce outage risk.

Critical customers such as hospitals may need additional backup even when the network is highly reliable.

The planning question is not whether centralized or decentralized systems are always more resilient. It is whether failure modes have been identified and whether the system can degrade gracefully.

Flooding belongs in the thermal-network map

Energy centres, pump stations and underground infrastructure can be vulnerable to flooding.

A network serving thousands of buildings should not place critical equipment at the lowest point of a flood-prone district without protection.

Climate-resilience planning should therefore shape plant elevation, access routes, electrical equipment placement and backup systems.

The thermal network is another example of why infrastructure design and The Climate Code cannot be separate conversations.

Customer protection matters because connection can create monopoly conditions

A building connected to one district network may have limited ability to switch suppliers.

That can create efficient infrastructure and market power.

Tariff rules, performance standards, transparency, metering, complaint processes and long-term governance therefore matter.

The exact regulatory model varies widely. Some systems are municipally owned. Some are utilities. Some are private concessions. Some are customer cooperatives.

The planner does not need to determine the tariff formula alone, but the spatial plan should not create mandatory network dependence without a credible consumer-protection framework.

Mandatory connection can help the network and harm a badly designed market

District systems sometimes use connection requirements because predictable demand supports financing.

The logic is understandable: if every new building can install its own competing system after the public sector has financed a shared network, the network may never reach viable load.

But mandatory connection needs conditions.

The network must offer reasonable cost, reliability and environmental performance. Exceptions may be needed where a building has a demonstrably better solution.

The policy should protect the shared investment without turning infrastructure planning into permanent technological lock-in.

Building readiness can be required before a network arrives

A district may be expected to receive district energy years after the first buildings are constructed.

Those buildings can be designed to connect later.

Plant-room space, pipe entry points, compatible internal temperatures and accessible heat-exchanger locations can reduce future retrofit cost.

Dublin’s project describes buildings in its catchment that are already “district heating enabled,” meaning they can connect when the network reaches them.

This is option value in building design.

Wastewater heat turns sanitation infrastructure into energy infrastructure

Wastewater leaves buildings warmer or cooler than the surrounding environment.

Large sewers and treatment plants can therefore serve as heat sources or sinks through heat exchangers and heat pumps.

The attraction is spatial: wastewater infrastructure often passes through dense urban areas where thermal demand is high.

But extraction must not compromise treatment processes, sewer operations or maintenance access.

TPW-0072 — The Sewer Capacity Map remains the wastewater-capacity owner. The thermal network simply recognizes that the same wastewater system can contain recoverable energy.

Cooling creates heat somewhere else

Air-conditioning removes heat from indoor space and rejects it outdoors.

At urban scale, this means cooling buildings can add heat to the surrounding environment.

District cooling can improve efficiency and concentrate heat rejection where it can be managed more effectively. In advanced networks, rejected heat may itself become useful.

The thermal planner therefore needs to think in flows: cooling is not destruction of heat. It is movement of heat.

Energy centres need real urban design

A district-energy plant is industrial infrastructure.

It can create noise, vents, loading, maintenance access and visual bulk.

In dense districts, plants may be placed inside mixed-use buildings, underground or beside other infrastructure.

The plant should therefore be planned as architecture and public-realm infrastructure, not hidden until the last design stage.

Singapore’s current bonus-GFA provisions for host developments are a useful example of planning explicitly for the space shared cooling infrastructure needs.

Land-value effects can support or obstruct the network

Shared energy infrastructure can increase development value by reducing plant-room requirements, improving environmental performance or providing reliable service.

It also imposes connection costs and may require easements or plant space.

Development agreements can allocate these costs and benefits.

If one host parcel provides a plant serving twenty surrounding buildings, it should not carry the land burden without compensation while everyone else receives the space benefit.

The financial architecture should follow the physical architecture.

Public land can de-risk the first energy centre

Early district networks often need a plant site before the market is mature.

Public land beside a civic complex, transit facility, wastewater plant or major redevelopment site can provide a strategic location.

The land does not need to be permanently dedicated if the plant is integrated into a larger development or designed for later adaptation.

The key is to avoid discovering after full build-out that the network has customers and no place for the infrastructure that serves them.

Procurement model changes the planner’s risk map

A thermal network can be developed by a municipality, utility, private energy-service company, development consortium or public–private partnership.

Each model allocates risk differently.

Who bears construction overruns? Who guarantees customer load? Who owns the pipes? Who pays when a source fails? Who funds expansion? Who maintains the street after repair?

The planning approval should understand these relationships because long-lived underground infrastructure can outlast the original developer.

Heat networks are especially sensitive to stranded-asset risk

A pipe network is expensive and difficult to relocate.

If the planned district never develops, if major customers disconnect or if the heat source closes, the network can become underused.

Scenario testing should therefore include weaker growth, lower thermal demand and source change.

Buildings are becoming more efficient. Climate change can alter heating and cooling demand. Technologies improve.

The network should not depend on permanently high demand merely because high demand makes today’s business case easier.

Efficiency can shrink the market while strengthening the system

If buildings improve insulation and equipment efficiency, each customer may need less thermal energy.

That can reduce revenue for a network designed around high consumption.

It is still the correct direction.

The business model should reward efficient service rather than depend on wasting energy.

Lower-temperature networks, larger service territories, new customers and diversified sources can help adapt to declining per-building demand.

Thermal zoning can reveal where different technologies fit

Not every neighbourhood should have district energy.

Brussels’ current approach to clean heating and cooling uses a zoned vision that maps which solutions fit which neighbourhoods, including district networks, heat pumps, geothermal systems and other sources.

This is a useful planning model.

Dense zones with large anchor loads may suit district networks. Lower-density areas may be better served by building-level heat pumps. Areas near industrial heat or wastewater sources may have another opportunity.

The town should choose technology by spatial condition instead of declaring one solution universally superior.

District cooling can release rooftop and basement space

Individual cooling systems require chillers, cooling towers, pumps and mechanical rooms.

Centralizing part of that equipment can free space in customer buildings.

That space can become usable floor area, rooftop solar, gardens, amenities or other building functions.

The planning value of a district system therefore includes land efficiency as well as energy efficiency.

This benefit should be included in feasibility analysis because urban land cost can materially change the economics.

Metering creates the accountability layer

A shared network needs to know who used what.

Building-level or customer-level metering supports billing, efficiency, fault detection and demand management.

Data can also reveal network performance: heat loss, peak demand, low-return-temperature problems, cooling efficiency and customer anomalies.

But metering data can become commercially and personally sensitive at high resolution.

The system should collect enough information to operate efficiently without creating unnecessary surveillance of household behaviour.

Digital optimisation is becoming part of thermal infrastructure

Modern networks increasingly use forecasting, sensors and optimization software to coordinate sources, storage and customer demand.

European district-energy programmes in 2026 are explicitly exploring digital tools for network design and real-time operations.

This can improve efficiency, but physical planning remains fundamental.

No algorithm can recover heat across a district if there is no pipe corridor, no customer connection and no legal right to cross the street.

Digital intelligence optimizes the network the town physically made possible.

A district-energy plan needs a source succession plan

Waste heat can disappear.

A factory closes. A data centre changes technology. A waste plant is decommissioned. A process becomes more efficient.

A network designed around one source should therefore identify the second and third sources before the first fails.

Large heat pumps, geothermal energy, electric boilers, thermal storage or additional waste-heat providers can create redundancy.

The long-lived asset should be the flexible network, not dependence on one industrial neighbour.

Heat-source contracts must last long enough to finance pipes

A customer may sign a ten- or twenty-year energy contract. A pipe can last much longer.

This mismatch matters.

If the network depends on a private waste-heat supplier, contracts should address source availability, maintenance outages, price, minimum delivery, expansion and closure.

Planning approval does not write the commercial contract, but planners should recognize that a thermal-source dot on a map is not secure until institutional arrangements make it reliable.

The public realm benefits when infrastructure is coordinated once

Repeated street excavation damages pavement, disrupts shops and creates construction fatigue.

A coordinated district can install thermal pipes alongside other utilities and reduce future digging.

Common services tunnels, utility corridors and synchronized capital works can make infrastructure more maintainable.

URA’s 2026 systems-level planning example is instructive because it treats district cooling and common utility distribution as parts of one spatial strategy rather than separate engineering projects.

Thermal networks can strengthen local energy security

Heating and cooling systems that depend heavily on imported fuels are exposed to price and supply shocks.

Local waste heat, geothermal energy, renewable electricity and thermal storage can diversify the energy mix.

The 2026 European district-heating outlook emphasizes this energy-security dimension as networks move toward renewable and recovered sources.

Local does not automatically mean resilient; local infrastructure can fail too.

The advantage is diversification: a city can use several sources that are difficult to integrate efficiently building by building.

Affordable housing should not become the captive customer paying for ambition

District-energy projects are often justified partly through climate goals.

If network tariffs become high, low-income residents can bear a disproportionate cost.

Social housing connections therefore need explicit affordability analysis.

Capital subsidy, tariff design, efficiency standards and customer protections can be used to ensure that decarbonisation does not increase energy poverty.

Brussels’ current Be.SHARE project is useful because it explicitly links clean district heating and cooling with affordability and includes social housing customers in the demonstration.

Waste heat has a justice geography too

Large industrial and data-centre facilities may create environmental burdens in neighbouring communities.

If public policy helps recover their waste heat, planners should ask who receives the benefit.

A network that sends useful heat only to premium commercial development while nearby lower-income housing bears infrastructure impacts can deepen inequity.

Connection strategy can therefore be part of the public-benefit agreement.

The system should not use “circularity” as a label while ignoring distribution.

Cooling demand will grow as heat risk grows

Climate change is increasing cooling needs in many cities.

That creates both a public-health challenge and an electricity-grid challenge.

District cooling, thermal storage, efficient buildings, shade and passive design should therefore be considered together.

A city should not use district cooling to excuse poor building design. Reducing cooling demand first makes the shared network smaller, cheaper and more resilient.

The Heat Refuge Network owns the emergency side of extreme heat. The Thermal Network owns the ordinary infrastructure that can reduce the energy cost of keeping buildings comfortable.

The thermal map should be updated as the district changes

A network planned from a one-time demand study can become obsolete.

Buildings convert use. Data centres arrive. Industry leaves. Efficiency improves. Climate changes. New technologies appear.

The town should therefore maintain a living thermal map that tracks major sources, loads, pipe capacity, connected customers, storage and future development.

This is not a public map of commercially sensitive operational data. It is a planning model of enough detail to guide infrastructure decisions.

A thermal-network feasibility screen

Before commissioning detailed engineering, planners can use a staged screen.

Density: Is enough thermal demand concentrated along a plausible route?
Anchors: Are there reliable large customers that can support the first phase?
Sources: Which renewable, recovered or conventional thermal sources are available?
Temperature: Does source quality match customer needs, directly or through heat pumps?
Timing: Do source availability and customer connection happen in compatible phases?
Corridor: Is there a continuous, maintainable route for pipes?
Plant: Is suitable land or building space available for energy centres and pumps?
Storage: Can thermal storage improve peak management or source utilisation?
Buildings: Are internal systems compatible, or can they become connection-ready?
Electricity: What grid capacity is required for pumps, chillers and large heat pumps?
Resilience: What happens when the main source, pump station or power supply fails?
Carbon: Is the real lifecycle emissions case better than credible alternatives?
Economics: Are capital cost, connection pace and customer tariffs viable?
Governance: Who owns, operates, regulates and expands the network?
Equity: Who bears infrastructure cost and who receives recovered-energy benefits?
Adaptability: Can the network accept future sources and lower building demand?

The Thermal Network in the wider Town Planning series

The thermal-network article sits between established owners without replacing them. The Hidden Town owns utility corridors. The Circular Town owns resource loops. The Data Centre District owns data-centre spatial impacts. The Time Layer owns infrastructure sequencing. The broader eduKate energy estate retains energy-system fundamentals.

The Thermal Network owns one spatial job: deciding when buildings are close, dense and complementary enough that sharing heat or cooling becomes better than every building solving the problem alone.

The city has always been an energy machine; the thermal network simply makes one hidden flow visible

Every cooled room pushes heat somewhere.

Every industrial process changes energy quality.

Every sewer carries thermal energy away. Every data centre creates a steady stream of rejected heat. Every building pays for comfort.

Town planning usually treats these as separate engineering facts.

The thermal-network perspective asks whether separation is wasting an opportunity.

Sometimes the answer is no. The buildings are too far apart, the loads are too small, the source is too uncertain, or individual systems are better.

Sometimes the answer is yes—and the difference is not a new appliance. It is an underground network, a shared institution and a town plan that recognized heat as something that moves through space.

That is how neighbouring buildings stop being separate energy islands and begin behaving like one urban thermal system.