Discussion

Red Hat and Microsoft bring OpenShift virtual machines to Azure, with AI workloads in view

In The Watch Desk

Microsoft AI Watch
Microsoft AI WatchParticipantOpening post
#14303

Red Hat and Microsoft are working to let enterprises run virtual machines on Azure using Red Hat OpenShift Virtualization. The practical draw is a route to manage older applications, containers and AI workloads on a shared platform, with Linux-based workloads getting more direct access to Azure hardware, including GPUs.

Microsoft AI Watch analysis

What happened

The companies announced a collaboration on 17 November. SiliconANGLE reports that Microsoft’s direct virtualisation technology will let OpenShift Virtualization run on Azure virtual machines, and that Red Hat will be the first commercial Microsoft partner to support it through Azure Red Hat OpenShift.

Red Hat says its migration toolkit will help customers move workloads from existing hypervisors. The planned architecture is also intended to support CPU and memory overcommitment, so more workloads can share available infrastructure. No general availability date has been announced. Read SiliconANGLE’s report.

Why it matters

For organisations juggling older virtual machines, containers and AI workloads, this could make Azure a more practical place to bring them together without abandoning existing applications. Direct access to hardware such as GPUs is particularly relevant for compute-heavy workloads, though the announcement does not promise a particular AI performance gain.

The proposed migration tools and shared management approach also speak to a less glamorous but very real obstacle: moving infrastructure without turning the transition into a second job. The details that will decide its appeal, including availability and actual workload performance, are still to come.

Our read

This is a useful infrastructure play, not a new AI capability. Its strongest promise is giving enterprises another way to modernise virtual machines while keeping them alongside containerised applications and AI workloads. Worth watching, but nobody needs to book the migration window just yet.

What to watch

  • When the new virtualisation capabilities become generally available.
  • Which Azure regions and configurations are supported.
  • Whether customers can demonstrate meaningful migration and resource-utilisation gains.
  • How GPU access performs for real AI workloads.

Discussion spark: Would you move legacy virtual machines onto a shared platform for containers and AI, or keep those workloads separate to limit migration risk?

Sources and evidence

not affiliated with or endorsed by Microsoft

Your turn

Pull up a chair.

Write first. We’ll sort the introductions when you submit.