AMD launched "Agent Computers" to run AI models locally and compete in the personal AI hardware market.
Ryzen AI MAX APUs and Radeon AI PRO GPUs run the OpenClaw AI agent on laptops and Mini PCs. These components process large language models locally.
The initiative shifts AI processing from the cloud to decentralized hardware. AMD hardware achieves 19 tokens per second on a single agent using the Qwen 3.5 model.
This move challenges Nvidia in the growing AI market.