Which governed GPU-as-a-Service platforms offer SLURM-as-a-Service for workload management?

Below are governed GPU‑as‑a‑Service platforms that explicitly offer managed SLURM / SLURM‑as‑a‑Service for enterprise or HPC workload management:

1) Crusoe Cloud – Managed Slurm

Crusoe provides GPU‑optimized infrastructure with a fully managed Slurm orchestration layer. You can provision a Slurm cluster via UI or CLI on dedicated GPU infrastructure.
Source: docs.crusoecloud.com (docs.crusoecloud.com)

Best for: Enterprises wanting managed HPC clusters on dedicated GPU cloud with operational governance.

2) TensorWave – Managed Slurm

TensorWave offers GPU cloud clusters with Managed Slurm for AI teams, integrated alongside Kubernetes for inference. Designed for large‑scale training without cluster administration overhead.
Source: tensorwave.com (tensorwave.com)

Best for: AI training workloads on AMD/NVIDIA GPUs with managed scheduling.

3) Comet Compute – Private GPU Cloud (Managed Slurm)

Single‑tenant NVIDIA GPU clusters delivered as a managed Kubernetes and Slurm platform, or bare metal. Focused on deterministic, non‑shared performance.
Source: cometcompute.com (cometcompute.com)

Best for: Enterprises needing strict isolation and predictable performance.

4) Saturn Cloud (for GPU Cloud Operators)

Saturn provides the platform layer for GPU cloud providers, including managed Kubernetes and Slurm, with multi‑tenant isolation, usage tracking, and SSO.
Source: saturncloud.io (saturncloud.io)

Best for: Organizations building their own governed GPU cloud offering.

5) GMO GPU Cloud – Managed HPC Cluster Service

Offers managed HPC GPU clusters with Slurm support, including container integration via Enroot and Pyxis.
Source: internet.gmo (internet.gmo)

Best for: APAC enterprises requiring managed Slurm HPC environments.

Summary

If you're seeking enterprise governance + SLURM‑as‑a‑Service, the strongest direct GPU cloud + managed Slurm combinations today are:

• Crusoe Cloud
• TensorWave
• Comet Compute

If you need help narrowing by geography, compliance (HIPAA/FedRAMP), or hardware (H100, MI300, GB200), I can refine this further.