Generative AI models brought about an influx of high-density cabinets. There has been much focus on how to best manage thermal issues, but the weight of power distribution equipment is a potentially overlooked concern.
Most operators will be familiar with the outrageous power and cooling demands of hardware for generative AI. Why are these systems so difficult to accommodate, and what does this mean for the future of data center design?
Douglas Donnellan
Andy Lawrence
Daniel Bizo
Peter Judge
John O'Brien
Jacqueline Davis
Max Smolaks
Jabari Williams-George
Rose Weinschenk
The 14th edition of the Uptime Institute Global Data Center Survey highlights the experiences and strategies of data center owners and operators in the areas of resiliency, sustainability, efficiency, staffing, cloud and AI.
Although there is still uncertainty around the rate of AI adoption, many organizations are pushing ahead to avoid being left behind. However, behind this enthusiasm, there are six issues that operators face when hosting AI.
Preliminary calculations by Uptime Intelligence suggest the initial impact of generative AI on global data center power use is low - but it will rise quickly as adoption increases. How far generative AI will go remains unclear.
Air-assisted direct liquid cooling systems offer trade-offs that make them attractive to operators looking to address server cooling and rack density challenges - and are relatively easy to install and maintain.
Uptime Intelligence looks beyond the more obvious trends of 2024 and identifies some challenging issues. Strong IT demand, high-density IT systems and tough sustainability requirements will drive a new wave of investment.
Despite high expectations, most operators will only see moderate impact from specialized AI hardware installations in the immediate future. The emergence of AI as a major force will sway the industry in a more profound, but less direct, fashion.
