Startups Tackle Data Centre Energy Crisis Amid AI Boom
Data centres account for roughly 2% of global electricity, which translates to around 1% of energy-related greenhouse gas emissions. The International Energy Agency predicts energy use from data centres will double by 2026, driven mainly by artificial intelligence. To combat this, startups are working on energy-efficient solutions.
One key solution is extracting the most value from every kilowatt. Startups like Asperitas, Submer, and Iceotope are working on immersion cooling, where servers are submerged in a non-conductive liquid that absorbs heat. This method can cut energy consumption of cooling a data centre in half.
Another approach is repurposing data centre heat. Startups like Deep Green and WINDCores are developing systems that use waste heat to power nearby facilities. For example, Deep Green’s system turns waste heat into hot water for swimming pools.
Balancing the energy demands of data centres with climate and infrastructure constraints is crucial. Governments and service providers must ensure data centres run on green power and operate efficiently. This might involve deploying data centres in space or powering them with fusion energy in the future.
While startups are making progress, there are challenges ahead. Tech giants are facing criticism for overinflating their progress on climate change. A recent investigation found that real emissions from company-owned data centres were higher than officially disclosed. To address this, governments and service providers must prioritize green power and energy efficiency in data centres.