This guide covers the RPM Fusion method — the recommended and most maintainable way to install NVIDIA proprietary drivers on Fedora. No manual .run installers, no broken dnf states.
Prerequisites
Before you start, make sure your system is fully up to date
sudo dnf upgrade --refresh
Reboot if there were kernel updates:
sudo reboot
Step 1 — Enable RPM Fusion Repositories
Fedora doesn't ship proprietary drivers by default. RPM Fusion fills that gap. You need both the free and nonfree repos:
sudo dnf install \
https://mirrors.rpmfusion.org/free/fedora/rpmfusion-free-release-$(rpm -E %fedora).noarch.rpm \
https://mirrors.rpmfusion.org/nonfree/fedora/rpmfusion-nonfree-release-$(rpm -E %fedora).noarch.rpm
## Step 2 — Install the NVIDIA Driver
Install the akmod-nvidia package. The akmod (automatic kernel module) system automatically rebuilds the driver whenever your kernel updates — saving you from the classic "driver broke after update" headache.
sudo dnf install akmod-nvidia
*Optional *
If you need CUDA support (for ML, video encoding, etc.), add this too:
sudo dnf install xorg-x11-drv-nvidia-cuda
Step 3 — Wait for the Kernel Module to Build
This is the step most guides skip over — and it's important.
After installation, the kernel module needs to compile. Do not reboot immediately. Wait for the build to complete:
sudo akmods --force
sudo dracut --force
You can verify the module is built by running:
modinfo -F version nvidia
You should see a version number like 570.xx.xx in the output. If you see an error instead, do not reboot yet — something went wrong.
Step 4 — Reboot
Once the module is confirmed built:
sudo reboot
After rebooting, verify the driver is active:
nvidia-smi
You should see your GPU listed with the driver version and memory usage.
Wrapping Up
The RPM Fusion + akmod-nvidia approach is the most reliable way to run NVIDIA drivers on Fedora. It survives kernel updates, works across Fedora versions, and doesn't require babysitting .run scripts.
Got questions or a specific GPU edge case? Drop them in the comments below. 🐧
