Firms turn from cloud to colo as they struggle with AI costs

Companies are grappling with rising energy costs and increased computing demands for AI workloads. Public cloud may not be the most cost-effective solution for deploying AI at scale. Some organizations are turning to colocation providers and GPU-as-a-service models as alternatives. IDC forecasts that spending on AI infrastructure will top $100 billion by 2028, with accelerated servers making up 56% of the market. However, growing concerns about energy consumption for AI could impact future datacenter expansion plans.