AI agents put CPUs back in demand for AMD and Intel

The rise of personal AI agents is creating fresh demand for CPUs after years in which Nvidia’s GPUs defined the generative AI boom. Agents that can plan tasks, operate software, browse the internet, access files, and run code need sustained computing environments, not just chips for model inference.

AMD and Intel are benefiting as cloud providers support agent workloads with virtual machines running on conventional processors. OpenAI’s Dots and Meta’s Muse rely on virtual computers, and Meta says it takes a CPU-agnostic approach to hardware. Investors have responded, with AMD and Intel shares rising 32% and 21%, respectively, over the past month.

AMD appears especially well positioned because its processors are widely used by major cloud providers, while its data-center revenue more than doubled to $6.7 billion in the quarter ended in June. Competition is also intensifying, with Arm, Nvidia, Intel, and cloud companies pursuing CPU opportunities tied to agentic computing. GPUs remain essential for training and inference, but CPUs are becoming more central to the infrastructure that lets agents work continuously in the background.