Home
Products
Red Hat AI Inference Server
3.1
Supported product and hardware configurations
Chapter 2. Supported AI accelerators

Chapter 2. Supported AI accelerators

The following tables list the supported AI accelerators for Red Hat AI Inference Server 3.1.

Important

Red Hat AI Inference Server 3.1 is not compatible with CUDA versions lower than 12.8.

Expand

Table 2.1. Supported NVIDIA AI accelerators
Container image	vLLM release	AI accelerators	Requirements	vLLM architecture support	LLM Compressor support
`rhaiis/vllm‑cuda-rhel9`	vLLM 0.9.0.1	Turing Ampere Ada Hopper Blackwell (compute capability 10.0)	CUDA Toolkit 12.8 NVIDIA Container Toolkit 1.14 NVIDIA GPU Operator 24.3 Python 3.12 PyTorch 2.7.0	x86 Aarch64 Developer Preview	x86 Technology Preview

Expand

Table 2.2. Supported AMD AI accelerators
Container image	vLLM release	AI accelerators	Requirements	vLLM architecture support	LLM Compressor support
`rhaiis/vllm‑rocm-rhel9`	vLLM 0.8.4	AMD Instinct MI210 AMD Instinct MI300X	ROCm 6.2 AMD GPU Operator 6.2 Python 3.12 PyTorch 2.7.0	x86	x86 Technology Preview

Expand

Table 2.3. Supported Google TPU AI accelerators
Container image	vLLM release	AI accelerators	Requirements	vLLM architecture support	LLM Compressor support
`rhaiis/vllm‑xla-rhel9`	vLLM 0.8.5	Google TPU v6e	Python 3.12 PyTorch 2.7.0	x86 Developer Preview	Not supported

Learn

Try, buy, & sell

Communities

About Red Hat Documentation

We help Red Hat users innovate and achieve their goals with our products and services with content they can trust. Explore our recent updates.

Making open source more inclusive

Red Hat is committed to replacing problematic language in our code, documentation, and web properties. For more details, see the Red Hat Blog.

About Red Hat

We deliver hardened solutions that make it easier for enterprises to work across platforms and environments, from the core datacenter to the network edge.

Theme

© 2026 Red Hat

Back to top