Air-Cooled AI Infrastructure: Agentic AI Drives CPU Demand

Air-cooled infrastructure is becoming a priority as agentic AI grows. While GPUs have dominated AI, CPUs are now gaining focus. This is because agentic AI relies heavily on constant orchestration and data processing, which are CPU strengths. Robert Hormuth from Advanced Micro Devices Inc. explains that agentic AI is goal-oriented. It uses various tools and software, including querying GPUs for complex math, but much of the work involves planning, checking, and iterating. This increased activity drives CPU demand. Dell Technologies Inc. is seeing a shift towards air-cooled AI infrastructure. Many enterprise customers are not ready for liquid cooling in their data centers. David Schmidt from Dell says that air-cooled systems are a pragmatic path forward. They allow AI deployment without the high cost and complexity of rebuilding. New servers offer a 13-to-1 consolidation ratio, freeing up power and rack space for agentic workloads. This matters because it shows how companies are making AI deployment more accessible and efficient within existing limitations.