AMD pushes local agentic AI with Ryzen AI Max PRO 400 Series

AMD is positioning its Ryzen AI Max and Ryzen AI Max PRO 400 Series processors as a platform for businesses looking to run agentic AI workloads locally instead of relying entirely on cloud infrastructure.

The new processors offer up to 16 Zen 5 CPU cores, a 55 TOPS neural processing unit, 40 GPU compute units and a large unified memory pool designed for local AI development and deployment.

Compared with the previous generation, the Ryzen AI Max 400 Series supports up to 192GB of total memory, with as much as 160GB available for GPU workloads. AMD said the Ryzen AI Max PRO 400 Series can run models with more than 300 billion parameters at 4-bit quantization without cloud offload.

AMD said local deployment can help companies keep sensitive data on-device while reducing dependence on recurring cloud inference costs.

The processors support development environments on Windows and Linux and work with AI tools and frameworks including PyTorch, vLLM, llama.cpp, Ollama, ComfyUI and LM Studio.

AMD is also pitching a hybrid approach, where businesses run suitable workloads locally while using cloud models for tasks that require more advanced capabilities.