Nvidia and Linux, a Question

Posted by mashally on Friday January 23, 2026 @10:27PM
from the self-post dept.

If most AI servers run on Linux and AI models require high-performance GPUs such as NVIDIA, yet NVIDIA GPUs are known to have issues on Linux, how do organizations handle this?

( 4 comments | on HN )

Re:Nvidia and Linux, a Question by OsrsNeedsf2P on Friday January 23, 2026 @10:56PM (#46740914)

In addition to Nvidia Linux issues mostly being centered around the desktop, proprietary modules for AI work is done well by Nvidia

[ Parent | on HN ]

Re:Nvidia and Linux, a Question by bigyabai on Friday January 23, 2026 @10:44PM (#46740861)

The issues mostly extend to desktop usage, and the majority of them are fixed now that Wayland has stabilized.

If you weren't using Nvidia hardware for desktop applications, the Linux issues were entirely negligible.

[ Parent | on HN ]

Re:Nvidia and Linux, a Question by __patchbit__ on Saturday January 24, 2026 @12:40AM (#46741354)

Tesla had their CEO sit in on an engineering meeting with nVidia to make progress. Do organizations have engineers to help nVidia when they get stuck?

[ Parent | on HN ]

Re:Nvidia and Linux, a Question by bigyabai on Saturday January 24, 2026 @07:12PM (#46749184)

Nintendo had their CEO sit in on Nvidia meetings to make progress. I don't think you quite understand what you're bragging about here.

[ Parent | on HN ]