AI Connect | Data Science Connect
AI ConnectExplore the 2027 portfolio

AI Hardware

AMD

CPUs, Instinct GPUs and adaptive chips for AI from data center to edge

Visit website ↗See if it’s a matchRequest an introduction

About AMD

AMD designs processors and accelerators for servers, PCs and embedded systems. Its Instinct GPUs target AI training and inference, supported by the open ROCm software stack, and it also sells EPYC server CPUs, Ryzen AI client chips and FPGA-based adaptive SoCs from the Xilinx acquisition.

Common use cases

  • LLM training and inference
  • HPC simulation
  • AI PCs
  • Embedded vision and robotics

Products and services

Instinct accelerators

Data center GPUs for AI and HPC.

EPYC processors

Server CPUs.

Ryzen AI

PCs processors with integrated NPUs.

Versal adaptive SoCs

FPGA-based chips for edge and embedded AI.

ROCm

Open software platform for GPU computing.

Recent product updates

  • ROCm 7.0 released in September 2025, adding MXFP4 support, an AITER operator library, and updated PyTorch/TensorFlow/JAX support.

Key features

  • Data center GPUs
  • Open ROCm software
  • Server CPUs
  • Adaptive SoCs
  • On-device NPUs
  • DPU networking

Integrations

  • PyTorch
  • Hugging Face
  • Dell
  • HPE
  • Supermicro
  • Microsoft Azure
  • Oracle Cloud

AMD alternatives

FAQs

How can I buy AMD AI hardware?

Mainly through server OEMs, system integrators and cloud providers that offer Instinct-based systems.

What software supports it?

ROCm, with support for PyTorch, JAX and other common frameworks.

Does AMD cover edge?

Yes, through Versal adaptive SoCs and Ryzen embedded products.

Work at AMD? Request updates to this profile. Update this profile

Compare with similar providers

All AI Hardware providers