AI is tripling energy, and data centers are facing cooling challenges

     Digitalisation, artificial intelligence (AI) and automation are intensifying the demand for data centre (DC) IT infrastructure, driving the need for effective cooling solutions. Increasing computing power generates more heat, which in turn requires a significant increase in cooling capacity to maintain performance, stability and equipment life.

     According to ABI Research, the number of data centers worldwide will more than double from 10,978 in 2023 to 2030. In 2023, DC operators consumed 300 terawatt hours of energy for cooling alone, a figure that is expected to triple by 2030, with a compound annual growth rate (CAGR) of 15%. The forecast highlights the significant challenges of powering and cooling old and new data infrastructure.

     The recent rejection of Google's data center in Dublin and regulatory action in Germany highlight the growing energy and sustainability challenges posed by data centers and the urgent need for renewable and carbon-free energy, as well as regulations that address energy, water, physical space and carbon emissions to mitigate the impact of data centers.

     "Cooling loads alone account for 40 percent of data center energy consumption. An effective cooling strategy requires a holistic, technology-independent approach to optimizing power use efficiency (PUE), water use efficiency (WUE), thermal management, and reducing operating costs. With the emergence of energy-intensive AI data centers, operators are under increasing regulatory pressure to adopt responsible computing practices, assess environmental impact, and leverage renewable energy sources." Rithika Thomas, senior analyst for sustainable technologies at ABI Research, explains.

     Regulations such as the Energy Efficiency Directive (EED), the European Data Center Code of Conduct (EU DC CoC), ASHRAE, ISO 50001 Energy Management System and the Singapore Green Data Center Standard are putting pressure on data center operators to regulate their energy, water and carbon footprints. According to Thomas, "Cooling costs are a significant indirect cost that operators have to face. A 'one size fits all' approach fails to take into account the impact of factors such as scale, location, infrastructure needs, cost, local regulations and workload that are critical to designing efficient cooling solutions."

     To stay ahead of the curve, data center operators are adopting hybrid and modular cooling technologies at the infrastructure level, These include solutions from companies such as Carrier Global Corporation, Danfoss, Daikin, Johnson Control, Rittal, Schneider Electric, STULZ GMBH, Trane and Vertiv, And device-grade liquid or immersion Cooling solutions from Green Revolution Cooling, Iceotope, LiquidStack or Submer.

     Today's data center cooling solutions offer high hardware efficiency, AI-driven monitoring, predictive maintenance, and seamless integration with DCIM/BAS systems to extend the life of IT equipment and reduce energy use. Lower operating costs, environmental impact reporting, carbon reduction and efficient cooling are essential for sustainable DC growth in the coming years.