Sweden's Data Centre Boom Could Rescue District Heating, but Only If Municipalities Demand It Before Permits Are Signed

• Data centres consume enormous electricity and generate waste heat that currently dissipates unused; feeding it into district heating grids could extend the economic life of urban heat infrastructure facing long-term demand decline.
• Svante Axelsson of Fossilfritt Sverige argues in Dagens Industri that siting decisions being made now will lock in infrastructure for decades — making this the window for strategic action.
• Without mandatory heat-recovery requirements or price signals, profit-maximising data centre operators have no incentive to invest in heat exchange systems, and municipalities lose all leverage once permits are granted.
• Sweden's district heating networks face shrinking demand as buildings become more energy-efficient, raising the question of who will pay for grid maintenance if new heat sources aren't integrated.

Sweden is in the middle of a data centre construction wave. Cheap electricity, grid capacity, political stability, and growing EU demand for data sovereignty have made the country one of Europe's most attractive locations for hyperscale facilities. Writing in Dagens Industri, Svante Axelsson — national coordinator of Fossilfritt Sverige (Fossil Free Sweden) — argues that this boom carries an overlooked opportunity: the enormous waste heat data centres produce could rescue Sweden's district heating networks, which face a slow structural decline as buildings become better insulated and less heat-hungry.

The physics are straightforward enough. A large data centre converts virtually all its electricity consumption into heat. That heat has to go somewhere. In most current installations, it dissipates into the atmosphere through cooling systems — energy produced, paid for, and thrown away. Meanwhile, Swedish district heating grids, built over decades at enormous public expense, are watching their customer base shrink. More efficient buildings need less heat. The revenue that keeps pipes maintained and pumps running is eroding. Axelsson's argument is that connecting data centres to district heating networks solves both problems: operators get a use for their waste product, and heating grids get a new, reliable, low-carbon heat source that could keep them economically viable for decades.

The logic is sound. The question Axelsson raises but does not fully answer is who will make it happen. Data centre operators choosing sites in Sweden are optimising for electricity price, grid connection speed, land availability, and proximity to fibre networks. Heat recovery requires additional infrastructure — heat exchangers, pipe connections, temperature-matching systems — that costs money and adds complexity. For a company building a facility to serve cloud customers in Frankfurt or Amsterdam, piping warm water to a Swedish municipal heating grid is not a core business concern. It is, at best, a nice-to-have that might generate modest revenue or local goodwill.

This is where the incentive structure gets interesting. Sweden is currently courting data centre investment without attaching significant conditions. The selling points are market-based: low electricity prices, renewable energy mix, cool climate reducing cooling costs. Municipalities competing for the jobs and tax revenue these facilities bring have every reason to make the permitting process as smooth as possible — and very little reason to add requirements that might send an operator to a competing site in Finland or Norway. Once a permit is granted and construction begins, the leverage disappears entirely. Retrofitting heat recovery into a facility designed without it is far more expensive than building it in from the start.

Axelsson frames this as a coordination problem requiring strategic foresight. That framing is generous. It is also a classic case of concentrated benefits and diffuse costs: the data centre operator captures the upside of cheap Swedish electricity while the municipality inherits the long-term cost of maintaining a heating grid that could have been sustained by waste heat but wasn't. Swedish municipalities have been here before — granting permits to large industrial facilities on favourable terms, only to discover years later that the conditions they failed to negotiate would have been worth billions.

The window is now, as Axelsson argues, precisely because siting decisions being made today will determine infrastructure configurations for twenty to thirty years. Every data centre built without heat-recovery capability is a facility that will dump usable energy into the sky for its entire operational life. Whether Swedish municipalities have the nerve to demand conditions from investors they are desperate to attract is another matter. Finland already requires waste heat recovery plans in several municipal energy strategies. Sweden, for now, is offering its electricity at market rates and asking for nothing in return.

Sources: Dagens Industri