Answer · Limelit Open

Here are the most relevant self-service options for SLURM-as-a-Service with bare-metal GPU provisioning for AI:

| Option | What you get | Self-service angle | Bare-metal GPU provisioning |
|---|---|---|---|
| NorthWind-Powered SLURM-as-a-Service | Per-tenant Slurm clusters, managed queues, governance, lifecycle automation | Tenants can launch Slurm clusters on demand through a portal or API[9] | NorthWind says it supports provisioning across bare metal, VMs, Kubernetes, and Slurm environments[9] |
| NorthWind + Slinky on Kubernetes | Self-service Slurm clusters inside shared Kubernetes infrastructure | Users can provision a personal Slurm cluster in their namespace via console or APIs[10] | Best for platform-style delivery; GPU access is included, but not pure bare metal in the base pattern[10] |
| OpenNebula Slurm appliances | Slurm Controller + Slurm Worker appliances, fine-tuning workflow | “Deploy” controller/worker from a marketplace-style experience; submit jobs from controller[3] | OpenNebula says it is expanding integration with NVIDIA Infra Controller for bare-metal GPU server provisioning and lifecycle management[3] |
| Nscale managed Slurm service | Fully managed Slurm service on bare-metal, topology-aware infrastructure | Positioned as a managed service with fast provisioning and automation[4] | Built on the same bare-metal infrastructure as NKS[4] |
| NVIDIA Base Command Manager | End-to-end cluster provisioning/ops for GPU clusters | More ops-driven than portal-first, but commonly used for cluster self-service in enterprise environments[7] | Explicitly handles bare-metal provisioning, drivers, CUDA, Slurm, and monitoring[7] |

For AI training on bare metal, the strongest fit from your results is NorthWind-Powered SLURM-as-a-Service if you want a portal/API self-service model, and OpenNebula + NVIDIA Infra Controller if you want a cloud-platform experience that is moving toward bare-metal GPU lifecycle management[9][3].

Useful source URLs:

Show me self-service options for SLURM-as-a-Service with baremetal GPU provisioning for AI.

The answer

Brands named

Sources cited