NVIDIA GPU not detected in ESXi host after installing drivers, VIBs

Products

• VMware vSphere ESXi

Issue/Introduction

After installing the physical GPU and installing the NVIDIA VIBs, the GPU is not detected by ESXi.   

A similar message may appear in the /var/run/log/vmkwarning.log:

2025-01-01T01:11:11.11Z cpu34:2150649)ALERT: NVIDIA: module load failed during VIB install/upgrade.

Running the nvidia-smi command will result in error:

[root@esxi01:~] nvidia-smi
Failed to initialize NVML: Unknown Error

Environment

ESXi 7.x
ESXi 8.x

Cause

This is likely the result of a faulty GPU or PCIe slot as the host did not detect the card when enumerating the available PCIe devices.

Resolution

This is a hardware issue and not a software issue.  Please consult with the appropriate hardware vendors for further assistance.  

Additional Information

If there is another working host with working GPU, attempting to swap the GPUs between the host may help confirm if the issue is with the GPU itself or the PCIe slot in the host system.