Canonical is adding AMD’s ROCm to the Ubuntu 26.04 LTS repos, making it faster for developers to install GPU-accelerated AI and HPC libraries for AMD Instinct and Radeon GPUs.
The news mirrors September’s NVIDIA CUDA announcement, when Canonical said it would package CUDA tools in the official Ubuntu repositories, negating multi-step installation hurdles for developers to get up and running with GPU-accelerated AI/ML tasks on Ubuntu.
Now, AMD already provide ROCm for Ubuntu directly via its own website/repo. What this news means is packaging and maintenance duties shift over to Canonical, which has created a dedicated engineering team to shoulder the task.
And yes: it’s committed to long-term maintenance of the packages on Ubuntu 26.04 LTS and says it will “also submit these packages for consideration in Debian”, which is the kind of share-the-love-fest we er, love to see.
ROCm (and related tools) will be a simple apt install rocm away on Ubuntu 26.04 LTS, and security updates and bug fixes will come through the usual update channels (no additional repos or manual dependency management required).
What is ROCm?
It’s AMD’s open-source software stack for GPU computing, providing runtimes and libraries needed to run machine learning frameworks like PyTorch and TensorFlow on AMD GPUs in data centers — but Radeon graphics cards are also supported.
It is a competitor to NVIDIA CUDA, but CUDA is the more widely used (and popular).
Currently, installing ROCm on Ubuntu requires adding AMD’s external repositories and manually managing dependencies. But in Ubuntu 26.04, ROCm installation is an apt command away.
Which is why AMD’s Andrej Zdravkovic says the move makes it “easier for developers and enterprises to deploy AMD solutions on supported systems” to run AI/ML workloads and high-performance computing apps across Ubuntu desktop, server, WSL, and Edge.
Most end-users won’t have need for this and, to be clear, it is NOT preinstalled by default. But the move treats AMD the same as NVIDIA, though whether easier installation for the former can help bolster its popularity against the latter remains to be seen.
More details on the Canonical blog.
Instagram Engineering
Shopify Engineering