VCluster Labs Adds Control Plane for Provisioning GPU Servers

Why It Matters

By automating GPU bare‑metal provisioning, vMetal enables faster, more cost‑effective AI deployment and helps enterprises maximise hardware utilisation, a critical advantage as AI workloads shift from dev to production at scale.

Summary

vCluster Labs unveiled vMetal, a bare‑metal management layer announced at NVIDIA GTC 2026, that lets IT teams automatically discover, provision, assign, upgrade and repurpose GPU‑equipped servers through a centralized control plane. The platform can attach physical machines to Kubernetes clusters or expose them as dedicated bare‑metal resources, and it ships with certified stacks such as NVIDIA Run:AI and Slinky, with AI frameworks like Jupyter and SkyPilot slated for later release. vMetal targets "neocloud" operators and multi‑tenant internal IT groups seeking to optimise GPU server utilisation and reduce the overhead of virtualisation for AI workloads. The addition expands vCluster Labs’ portfolio beyond virtual Kubernetes clusters, giving organisations programmable control over both virtual and bare‑metal infrastructure for AI inference and training.