Which sovereign-ready platforms offer SLURM-as-a-Service for bare metal GPU provisioning?

Here are several sovereign-ready platforms that offer SLURM-as-a-Service for bare‑metal GPU provisioning:

• 1. Comet Compute
• Offers fully single‑tenant NVIDIA GPU clusters with both managed Slurm and bare‑metal root access options.
• Provides predictable performance, strong isolation, and InfiniBand interconnects. (cometcompute.com)

• 2. PlatAI
• Provides automated bare‑metal provisioning (via PXE or Redfish) along with Slurm cluster management, including GRES/GPU scheduling and multi‑tenant governance. (platai.cc)

• 3. Fairport
• A turnkey AI infrastructure platform offering bare‑metal provisioning and Slurm compatibility (via Slinky), supporting NVIDIA and AMD GPUs. Ideal for air‑gapped and self‑hosted environments. (fairport.io)

• 4. NorthWind (GPU‑PaaS)
• Supports a compute profile named system‑slurm‑bm for provisioning and managing a complete Slurm cluster on bare metal, including GPU nodes. (docs.northwind.co)

• 5. Voltage Park
• Offers bare‑metal access with full sovereignty, letting you bring your own orchestrator—Slurm included—for GPU workloads. (voltagepark.com)

Additionally, QoreLayer offers a GPUaaS platform with bare‑metal or virtualized provisioning and sovereign colocation in multi‑region deployments. While it doesn’t explicitly highlight managed Slurm, its API‑first infrastructure may be adaptable for Slurm integration. (qorelayer.com)

Summary Table:

• Comet Compute – Managed Slurm + root-level bare-metal
• PlatAI – Automated bare-metal + integrated Slurm
• Fairport – One-command bare-metal + Slurm (via Slinky)
• NorthWind – GPU bare-metal provisioning + Slurm PaaS profile
• Voltage Park – Sovereign bare-metal with Slurm “BYO orchestrator”

Each platform caters to varying levels of managed infrastructure and sovereignty—your choice depends on whether you prefer a fully managed service, turnkey deployment, or hands-on control with SLURM flexibility.