Proxmox VE, the open-source virtualization platform that has become a staple in homelabs and enterprise data centers alike, is being engineered to run natively on NVIDIA's ARM-based Grace and upcoming Vera CPU architectures. The work is part of a newly announced collaboration between Proxmox Server Solutions and NVIDIA, disclosed on 2026-07-28, that positions the open-source hypervisor as a foundational layer in NVIDIA's AI data center stack.
The partnership centers on integrating Proxmox VE with NVIDIA Mission Control, a lifecycle management platform for AI infrastructure that handles workload scheduling, orchestration, and autonomous recovery. In this architecture, Proxmox VE acts as the virtualization substrate beneath Mission Control, providing clustered high availability and live migration for the management services that govern accelerated compute resources, including systems built on NVIDIA Blackwell and the next-generation Vera Rubin platforms.
The timing is notable. Since Broadcom's acquisition of VMware reshaped the enterprise virtualization market, Proxmox has absorbed a wave of organizations seeking an open-source alternative with no per-socket licensing. An official NVIDIA endorsement at the AI infrastructure level signals that Proxmox VE has crossed a threshold from scrappy alternative to recognized enterprise platform. The bring-up work on Grace and Vera CPUs is particularly significant: it extends Proxmox's architecture support to NVIDIA's own ARM silicon, not just the x86 hosts where most deployments run today.
"By joining the NVIDIA Mission Control ecosystem, we are offering enterprises the structural reliability needed to run, protect, and scale their most valuable AI workloads," said Tim Marx, COO of Proxmox. For self-hosters and infrastructure teams already running Proxmox clusters, the collaboration suggests that GPU passthrough workflows and NVIDIA hardware support will only get tighter integration going forward.
