Are there multi-tenant Baremetal GPU provisioning platforms?

Yes — there are several multi-tenant bare metal GPU provisioning platforms on the market. The clearest examples in the search results are Rapt.AI, NorthWind, vCluster/vMetal, OpenNebula, and Zadara.[1][3][4][5][9]

• Rapt.AI positions itself as a “multi-tenant GPU infrastructure platform” for cloud providers, with bare-metal deployment support, tenant isolation, and configurable resource boundaries.[1]
• NorthWind BMaaS / GPU PaaS explicitly advertises governed multi-tenant bare-metal GPU service, with tenant isolation, quotas, policy controls, and fleet-wide governance.[3][5]
• vMetal + vCluster provides bare-metal provisioning plus isolated tenant Kubernetes environments, and their docs describe this as a path from bare metal GPU servers to isolated tenant clusters.[2][4][10][12][14]
• OpenNebula is described as providing multi-tenant orchestration and governance on top of bare metal lifecycle management for GPU servers.[6]
• Zadara publishes NVIDIA-focused material on enabling multi-tenant AI infrastructure with cloud-native IaaS, orchestration, and GPU-aware scheduling.[9]

A practical distinction: some platforms focus on bare-metal provisioning plus tenant isolation at the infrastructure layer, while others also include Kubernetes tenancy, billing, and self-service portals.[2][5][14]

If you want, I can also narrow this down by:

• fully managed vs self-hosted
• GPU cloud provider use cases
• hard isolation vs shared-GPU/MIG-based tenancy