Berlin’s data center boom is creating an unusual energy opportunity. Every server turns electricity into heat. Until recently, operators mainly saw that heat as something their cooling systems needed to remove. Berlin now has a reason to look at it differently. The city’s 2026 heat plan maps waste-heat sources against local heat demand. At the same time, Stromnetz Berlin expects peak electricity demand to rise from about 2 GW today to 4.5 GW by 2035. Data centers will contribute to that growth. The question is whether Berlin can turn some of that new demand into useful heat.
Berlin’s data centers are becoming permanent heat infrastructure
Data centers produce heat throughout the year. That makes them very different from many renewable heat sources.
The heat source already exists inside Berlin’s digital infrastructure
A data center does not need a separate process to create usable heat. Its servers, power systems and cooling equipment already release large amounts of thermal energy. The challenge lies in collecting that heat at a useful temperature. However, data centers have one important advantage. They can provide heat with far less weather dependence than solar thermal systems. That consistency matters to Berlin heating because homes, offices, and public buildings need a reliable supply. The real opportunity therefore sits around the facility. A data center surrounded by buildings with suitable heat demand has a very different value from one sitting far from any network.
Berlin’s heat plan makes location a strategic question
Berlin has been unusually upfront about waste heat. The city’s 2026 heat-planning data includes known industrial and commercial sources. It also shows heat-consumption density, existing networks and potential areas for local heating systems. So planners can align a heat source with actual demand at the neighborhood level. That’s important to keep in mind: A large source of heat is useless if there’s nobody close enough to use it. Berlin heating decisions can now be guided by a much more realistic question. Where is the recoverable heat located relative to the concentrated demand? Berlin’s data also include estimates of daily availability and thermal output. Still, site-level engineering and commercial diligence is vital.
The real constraint is matching heat, not finding it
Berlin has waste heat. The harder job involves matching its temperature, distance and timing with a network that can use it.
Low-temperature heat creates a second engineering problem
Data centers regularly reject heat at temperatures lower than those required for conventional district heating. Spandau (Berlin) project of NTT DATA serves as a straightforward example. Its two data centers generate heat at about 20–30°C. The local network to be realized requires about 65°C. A heat pump, on the other hand, must increase the temperature before it can be utilized by the network. That process consumes electricity and adds equipment. Instead, a low-temperature network can lower the temperature lift between neutrals. But the economics hinge on that delta. Data center cooling is also important, since cooling architecture impacts the temperature that can be recovered. So a worthwhile waste-heat project begins with engineering and not with a headline number for the amount of recoverable heat.
Schöneberg and Spandau show two very different opportunities
Berlin already has two promising projects. In Schöneberg, Deutsche Telekom transports heat from a data center to the Pallasseum. A water-to-water HP (heat pump) increases the temperature up to approximately 70–75 °C. The system services about 500 flats and 2,000 people. NTT DATA is planning to send up to 8MW from two existing data centers via a 2-kilometre pipeline in Spandau. It is the first to supply heat to Neue Gartenfeld, a new development featuring 4,500 apartments and 200 commercial spaces. The contrast matters. One system connects an existing building to an existing facility. The other links an existing heat source to a new district.
Berlin’s power grid could decide how far the model goes
Waste heat can reduce separate heat-generation needs. It cannot remove the electricity demand that data centers and heat pumps create.
Data centers and heat pumps will compete for connection capacity
Stromnetz Berlin expects peak demand to reach about 4.5 GW by 2035. The operator points to large data centers, heat pumps, new districts and electric mobility as major drivers. Meanwhile, Berlin has already introduced a separate allocation process for very large grid connections above 3.5 MW. The process specifically covers projects such as data centers and large heat pumps. That creates an important link between Berlin district heating and electricity infrastructure. A recovered-heat project still needs pumps, controls, and circulation equipment. Consequently, planners must look at the complete power requirement. Waste heat can lower the need for separate heat generation. It cannot make the data center electrically invisible.
A data center site should now carry an energy profile
Developers normally judge a data center site through power, fibre connectivity, land, cooling and resilience. That list needs another question. What happens to the heat? A site near dense housing or a commercial district can create a very different business case from an isolated facility. More importantly, developers should examine how demand will change over time. A housing district may need large amounts of heat in winter. A hospital, university, or industrial customer may create a steadier profile. The same logic applies to the electricity side. A site needs enough grid capacity for computing and heat-pump demand. Berlin heating could therefore become part of future data-center site selection.
Berlin could turn digital growth into a heating advantage
The biggest opportunity may sit beyond existing facilities. New districts give Berlin a chance to plan digital and heating infrastructure together.
New districts can plan around predictable digital heat
Neue Gartenfeld shows what this approach can look like. The development will include homes, commercial units, schools and childcare facilities. NTT DATA plans to supply up to 8 MW of heat from its nearby data centers. A local energy centre will distribute that heat through a neighbourhood network. The project also includes a 300-cubic-metre hot-water tank and a 3.6 MW power-to-heat boiler for peak winter demand. In practice, the project does not depend on a single source. It combines recovered heat, storage and backup capacity. That makes waste heat recovery part of the system design rather than an extra feature added after construction.
Berlin can turn heat mapping into an investment tool
The city’s new heat maps can do more than support public planning. Developers can use them when screening future sites. Utilities can identify areas where a heat source could support a new network. Data-center operators can identify nearby customers before they commit to a location. Likewise, property developers can consider future data centers when planning local energy systems. This creates a more interesting model for waste heat in Berlin. Instead of asking whether an existing facility can sell surplus heat, planners can ask where a future facility could create the most value. Berlin heat network planning then becomes part of investment strategy. That could change how developers view energy infrastructure across Berlin.
To sum up
Berlin’s data centers should not be judged only by how much electricity they consume. They can also provide a steady local heat source. Yet that opportunity depends on location, temperature, network design and grid capacity. Berlin heating can gain real value when planners consider those factors before new infrastructure becomes fixed. The 3rd Industrial Energy Infrastructure, Power & Thermal Systems Summit takes place on 10–11 September 2026 in Berlin, Germany. It brings together industrial energy professionals around grid capacity, thermal systems, storage, waste heat, and infrastructure planning. Its practical focus makes it especially relevant to teams shaping Europe’s next industrial energy systems.



