43 Kinoite on Framework 16, nvidia drivers are installed and detected, but DGPU is not activated and fullscreen DGPU applications only show black screen

See in the title, I’m currently running fedora 43 Kinoite on a framework 16 with a Ryzen AI 9 HX 370 Mainboard and RTX 5070 GPU module.

Sometimes when I boot, the DGPU appears to remain in idle mode, even when an application requests it, leading to any applications that use the DGPU for rendering (primarily games and the godot engine editor in my use case) showing black screens if fullscreen, or failing to start in other cases.

However, the GPU driver does appear to still load even in this case, nvidia-smi output:

Sun Apr 19 12:11:30 2026       
+-----------------------------------------------------------------------------------------+
| NVIDIA-SMI 580.126.18             Driver Version: 580.126.18     CUDA Version: 13.0     |
+-----------------------------------------+------------------------+----------------------+
| GPU  Name                 Persistence-M | Bus-Id          Disp.A | Volatile Uncorr. ECC |
| Fan  Temp   Perf          Pwr:Usage/Cap |           Memory-Usage | GPU-Util  Compute M. |
|                                         |                        |               MIG M. |
|=========================================+========================+======================|
|   0  NVIDIA GeForce RTX 5070 ...    Off |   00000000:C1:00.0 Off |                  N/A |
| N/A   41C    P8              5W /  100W |       2MiB /   8151MiB |      0%      Default |
|                                         |                        |                  N/A |
+-----------------------------------------+------------------------+----------------------+

+-----------------------------------------------------------------------------------------+
| Processes:                                                                              |
|  GPU   GI   CI              PID   Type   Process name                        GPU Memory |
|        ID   ID                                                               Usage      |
|=========================================================================================|
|  No running processes found                                                             |
+-----------------------------------------------------------------------------------------+

Sometimes, uninstalling and reinstalling akmod-nvidia, xorg-x11-drv-nvidia and xorg-x11-drv-nvidia-cuda from rpmfusion and rebooting fixes it for that boot specifically, but this is inconsistent and not always the case.

Attempting to use akmods --rebuild --force is similarly inconsistent, the outcome is included here, but I have assumed the installation failure is due to being on the atomic spin:

Checking kmods exist for 6.19.12-200.fc43.x86_64 [  OK  ]
Building and installing nvidia-kmod [FAILED]
Could not install newly built RPMs. You can find them and the logfile in:
/var/cache/akmods/nvidia/580.126.18-1-for-6.19.12-200.fc43.x86_64.failed.log
 [FAILED]

Bios DGPU setting: Hybrid

rpm-ostree status:

State: idle
Deployments:
● fedora:fedora/43/x86_64/kinoite
                  Version: 43.20260417.0 (2026-04-17T00:31:29Z)
               BaseCommit: a1dc5735c3d17604d4d444d253527100890f57f1157927a109dfa1865f16f8d8
             GPGSignature: Valid signature by C6E7F081CF80E13146676E88829B606631645531
          LayeredPackages: akmod-nvidia akmods distrobox libvirt lm_sensors rpmdevtools steam-devices tailscale virt-manager
                           xorg-x11-drv-nvidia xorg-x11-drv-nvidia-cuda
            LocalPackages: rpmfusion-free-release-43-1.noarch rpmfusion-nonfree-release-43-1.noarch

  fedora:fedora/43/x86_64/kinoite
                  Version: 43.20260417.0 (2026-04-17T00:31:29Z)
               BaseCommit: a1dc5735c3d17604d4d444d253527100890f57f1157927a109dfa1865f16f8d8
             GPGSignature: Valid signature by C6E7F081CF80E13146676E88829B606631645531
          LayeredPackages: akmod-nvidia akmods distrobox libvirt lm_sensors rpmdevtools steam-devices tailscale virt-manager
                           xorg-x11-drv-nvidia xorg-x11-drv-nvidia-cuda
            LocalPackages: rpmfusion-free-release-43-1.noarch rpmfusion-nonfree-release-43-1.noarch

Output of rpm-ostree kargs

ostree=/ostree/boot.0/fedora/0d68e87a50b38221cd5905c9f21bac44ce38a4b6c1e16d8c1dbe68fd4640a01f/0 rhgb quiet root=UUID=a54568e6-e38e-4e98-b5d9-e3dca31576df vconsole.keymap=us rootflags=subvol=root rw rd.driver.blacklist=nouveau,nova_core modprobe.blacklist=nouveau,nova_core

Output of dmesg | grep -iE 'secure|nvidia|nouveau'

[    0.000000] Command line: BOOT_IMAGE=(hd0,gpt2)/ostree/fedora-0d68e87a50b38221cd5905c9f21bac44ce38a4b6c1e16d8c1dbe68fd4640a01f/vmlinuz-6.19.12-200.fc43.x86_64 ostree=/ostree/boot.0/fedora/0d68e87a50b38221cd5905c9f21bac44ce38a4b6c1e16d8c1dbe68fd4640a01f/0 rhgb quiet root=UUID=a54568e6-e38e-4e98-b5d9-e3dca31576df vconsole.keymap=us rootflags=subvol=root rw rd.driver.blacklist=nouveau,nova_core modprobe.blacklist=nouveau,nova_core
[    0.000000] secureboot: Secure boot disabled
[    0.003332] secureboot: Secure boot disabled
[    0.025103] Kernel command line: BOOT_IMAGE=(hd0,gpt2)/ostree/fedora-0d68e87a50b38221cd5905c9f21bac44ce38a4b6c1e16d8c1dbe68fd4640a01f/vmlinuz-6.19.12-200.fc43.x86_64 ostree=/ostree/boot.0/fedora/0d68e87a50b38221cd5905c9f21bac44ce38a4b6c1e16d8c1dbe68fd4640a01f/0 rhgb quiet root=UUID=a54568e6-e38e-4e98-b5d9-e3dca31576df vconsole.keymap=us rootflags=subvol=root rw rd.driver.blacklist=nouveau,nova_core modprobe.blacklist=nouveau,nova_core
[    4.244277] amdgpu 0000:c2:00.0: amdgpu: SECUREDISPLAY: optional securedisplay ta ucode is not available
[    8.247859] nvidia: loading out-of-tree module taints kernel.
[    8.247870] nvidia: module verification failed: signature and/or required key missing - tainting kernel
[    8.376150] nvidia-nvlink: Nvlink Core is being initialized, major device number 509
[    8.378667] nvidia 0000:c1:00.0: enabling device (0000 -> 0003)
[    8.378802] nvidia 0000:c1:00.0: vgaarb: VGA decodes changed: olddecodes=io+mem,decodes=none:owns=none
[    8.390470] NVRM: loading NVIDIA UNIX Open Kernel Module for x86_64  580.126.18  Release Build  (dvs-builder@U22-I3-H04-01-6)  Wed Feb 11 18:33:27 UTC 2026
[    8.456406] nvidia-modeset: Loading NVIDIA UNIX Open Kernel Mode Setting Driver for x86_64  580.126.18  Release Build  (dvs-builder@U22-I3-H04-01-6)  Wed Feb 11 18:19:14 UTC 2026
[    8.493533] [drm] [nvidia-drm] [GPU ID 0x0000c100] Loading driver
[    8.982560] nvidia 0000:c1:00.0: Enabling HDA controller
[    9.002767] input: HDA NVidia HDMI/DP,pcm=3 as /devices/pci0000:00/0000:00:03.1/0000:c1:00.1/sound/card2/input26
[    9.002991] input: HDA NVidia HDMI/DP,pcm=7 as /devices/pci0000:00/0000:00:03.1/0000:c1:00.1/sound/card2/input27
[    9.003018] input: HDA NVidia HDMI/DP,pcm=8 as /devices/pci0000:00/0000:00:03.1/0000:c1:00.1/sound/card2/input28
[    9.003091] input: HDA NVidia HDMI/DP,pcm=9 as /devices/pci0000:00/0000:00:03.1/0000:c1:00.1/sound/card2/input29
[   11.450441] [drm] Initialized nvidia-drm 0.0.0 for 0000:c1:00.0 on minor 0
[   11.450896] nvidia 0000:c1:00.0: [drm] Cannot find any crtc or sizes

I don’t know if this is an issue similar to mine, but I’m running Fedora 43 KDE on a laptop with Intel iGPU + Nvidia 2050 and games always launch with a black screen after a fresh startup.

The only workaround I found is launching a Vulkan app and forcing the Nvidia driver to be used before I play any game:

VK_DRIVER_FILES=/usr/share/vulkan/icd.d/nvidia_icd.x86_64.json vkcube

After the application launches, I close it immediately and then everything else works fine.

Thanks, the GPU is working as expected currently, but next time this happens I will try it