AI Infrastructure

AI infrastructure encompasses the GPUs, memory, optical communications, power, cooling systems and data centers that support model training and inference. Continued growth in generative AI investment has recently driven up memory prices, demand for optical modules and deployments of liquid-cooled racks. Countries buying more GPUs are also facing pressure on data-center space and electricity supplies, while cybersecurity threats targeting computing resources are rising. The supply chain affects semiconductors, energy costs and shipments of end-user devices, making shifts in capacity, pricing and investment returns important areas to watch.