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

Below are sovereign-ready platforms that provide (or explicitly support) SLURM-as-a-Service on bare‑metal GPU infrastructure, suitable for regulated/public-sector or EU data‑sovereignty use cases:

1) PeakOps – Managed Slurm on Sovereign / On‑Prem HPC

What you get: Fully managed Slurm (cluster health, scheduling, QoS, capacity planning) on your own bare‑metal GPU hardware or sovereign environments. Designed for air‑gapped and regulated deployments.
Why it fits: True Slurm‑as‑a‑Service with sovereignty control (runs in your DC or sovereign setup).
Source: peakops.co (peakops.co)

2) Oreus Cloud (France) – Sovereign GPU Bare Metal + Slurm

What you get: French/EU sovereign GPU cloud with bare‑metal GPU, Kubernetes, Docker, and Slurm support. Operated in France/EU for compliance‑sensitive workloads.
Why it fits: Explicit sovereign positioning + Slurm + bare‑metal GPU.
Source: oreus.ai (oreus.ai)

3) Polarise / Nscale (EU Sovereign AI Factories)

What you get: Sovereign AI infrastructure across Europe including bare‑metal GPU, managed Kubernetes, and Slurm. API-first, GDPR‑aligned platform.
Why it fits: Built specifically as sovereign AI factories with managed Slurm environments.
Source: vcluster.com (vcluster.com)

4) Voltage Park – Bare Metal GPU + Slurm Support

What you get: Direct bare‑metal GPU access with freedom to use Slurm, Kubernetes, Ray, etc.
Why it fits: Not fully managed Slurm by default, but supports Slurm on dedicated hardware; suitable for sovereign or single‑tenant deployments.
Source: voltagepark.com (voltagepark.com)

5) Fairport – Full‑Stack AI Infra (Slurm Compatible)

What you get: Platform covering bare‑metal provisioning through AI inference, with built‑in Slurm compatibility.
Why it fits: Slurm‑compatible sovereign-ready AI stack with infrastructure control.
Source: fairport.io (fairport.io)

Buyer Guidance

If you require:

• Fully managed Slurm operations → Start with PeakOps
• EU sovereign cloud region + Slurm baked in → Oreus or Polarise/Nscale
• Single-tenant bare metal with scheduler freedom → Voltage Park

If you’d like, tell me:

• Required region (US/EU/other)
• GPU type (H100, B200, MI300, etc.)
• Air‑gapped vs hosted sovereign
• Expected node count

I can narrow this to 2–3 best-fit vendors and suggest an RFP checklist.