Generative AI: Implications for the Datacenter

Generative AI: Implications for the Datacenter

This IDC Market Perspective provides insights into how the demanding energy requirements necessary to train generative AI models will change datacenter design, power distribution, cooling, and location. Because of its size, complexity, and drive toward sustainability, the market is likely to be dominated by a handful of cloud service providers, select OEMs, and niche cloud and colocation providers."Due to the intensive energy demands for training generative AI Models, datacenters performing these tasks will be purpose built, resembling other high-performance compute environment with high power density racks and liquid cooling, but they will increasingly be placed in locations outside of key metropolitan areas to take advantage of lower energy costs and free air cooling and avoid power capacity constraints," says Sean Graham, research director, IT Datacenters at IDC.

Please Note: Extended description available upon request.