As AI data centres guzzle energy, can we harness their wasted heat?

The rapid expansion of AI data centres is straining global energy infrastructure and exacerbating the world's energy crisis. These centres consume vast amounts of electricity, water, and are likely contributing to global warming. In 2024, data centres used 485 terawatt hours of electricity, the equivalent of 1.5 percent of the world's total energy supply.

Danish engineers have developed a solution to harness the waste heat generated by these data centres. The Bitten supercomputer at the University of Southern Denmark's Sonderborg campus recovers 100% of its waste heat, which was previously lost to the environment. This heat can then be used to warm Sonderborg's 75,000 residents through the town's municipal district heating network.

Danfoss, a Danish engineering company, developed advanced liquid cooling technology for the Bitten supercomputer. This system absorbs the waste heat from the high-temperature processors, which is then channelled into the town's district heating network. This innovation allows the waste heat to be repurposed for heating homes and industrial processes, rather than being released into the atmosphere.

Scandinavia has been at the forefront of capturing waste heat from data centres. Meta's data centre in Odense generates enough waste heat to power 6,900 homes annually. Stockholm has been feeding data centre heat into its network since 2014, and other potential sources of waste heat, such as supermarket fridge systems and factories, are also being explored.

Danfoss believes that this technology could help ease the increasing demand for electricity caused by the AI boom, reducing reliance on fossil fuels and making business sense. As more data centres come online, the scalability of this waste-heat technology could play a significant role in easing the strain on global energy supply.