- Published on
Can Europe Build Data Centers Without Giving Up Its Climate Goals?
- Authors
- Name
- DeepHorus
Germany’s push to expand AI data centers is colliding with climate commitments after Berlin weakened rules requiring new facilities to reuse waste heat. The amended Energy Efficiency Act, due before parliament imminently, will require operators only to assess whether heat recycling is feasible, not to do it. The retreat follows industry complaints that the original 2023 rules made projects nearly impossible.
Germany aims to double its data center fleet by 2030 and quadruple AI compute, reducing U.S. dependence and boosting its economy. But EU states must cut greenhouse gas emissions by 55 percent by decade’s end. The original law sought to recycle waste heat into municipal heating, which former official Patrick Graichen called a “no-brainer.”
Firms argued the obligation was unworkable when no heating system or industry could take the heat. The German Data Center Association called an early version a “data center prevention act.” Barbara Smitten of Cundall said some requirements were outside operators’ control.
The uncertainty has already hit projects. In the last 18 months, Stack Infrastructure canceled a $3 billion facility in Hesse, and state-owned Komm.ONE withdrew from a Baden-Württemberg project citing the law’s burdens. The German Ministry for Economic Affairs and Energy did not respond to a request for comment.
Graichen accused the government of sacrificing climate targets for the AI wave. He said heat offtake access should be a default siting criterion and told the industry to “stop complaining.” John Booth of Carbon3IT said no technical reason prevents data centers from reusing waste heat; the technologies are known and only organizational will is missing.
The amended bill faces parliament amid European tension between energy-intensive AI growth and climate goals. It requires feasibility assessments and exempts operators when recycling is not viable. The dispute shows the difficulty of building AI infrastructure without undermining emissions targets.
