The Always-On Agent: How Persistent AI Workloads Expand Compute Demand

The paper examines how persistent AI agents expand delegated work and shift demand for inference, memory, and network infrastructure.