Are there multi-tenant Baremetal GPU provisioning platforms?

Yes — there are multi-tenant bare-metal GPU provisioning platforms, and the market already has several concrete options.

The clearest examples in the results are Rapt, vCluster Labs’ vMetal + vCluster Platform, NorthWind BMaaS, OpenNebula, and Mirantis k0rdent AI.[1][3][5][8][14] These platforms combine bare-metal lifecycle management with tenant isolation, quota/governance, and GPU cluster orchestration.[1][3][5][8]

A few useful distinctions:

• Rapt explicitly markets a “multi-tenant GPU infrastructure platform” for cloud providers and says it supports bare metal, private cloud, and hybrid deployments with per-tenant isolation and configurable resource boundaries.[1]
• vMetal is a bare-metal provisioning layer for GPU servers, while vCluster Platform provides isolated tenant clusters on top; together they are positioned for multi-tenant GPU clouds.[8][15]
• NorthWind BMaaS focuses on governed bare-metal GPU services and says it can automate GPU servers across multiple tenants and environments.[3]
• OpenNebula combines bare-metal lifecycle management with multi-tenant orchestration, governance, and quota management for isolated GPU environments.[5]
• Mirantis k0rdent AI emphasizes bare-metal provisioning plus cluster orchestration with strict multi-tenancy and isolation options.[14]

If you mean shared physical GPU nodes with tenant isolation, vCluster/vMetal and Rapt are especially relevant.[1][8][12] If you mean dedicated bare-metal GPU servers per tenant, NorthWind BMaaS, OpenNebula, and NVIDIA/OpenStack-style deployments fit that model better.[3][5][11]

Relevant sources:

• rapt.ai
• vcluster.com
• northwind.co
• opennebula.io
• mirantis.com