Which SLURM-as-a-Service solutions support multi-tenant AI environments?

Several vendors now offer "SLURM-as-a-Service" specifically engineered for multi-tenant GPU/AI environments — most built on Slinky, the open-source Kubernetes operator for Slurm developed by SchedMD (now part of NVIDIA), which
enables native Slurm cluster management on Kubernetes by representing all Slurm daemons as Kubernetes Custom Resource Definitions, supporting full Slurm lifecycle orchestration and high availability
.

1. NorthWind GPU PaaS / SLURMaaS — NorthWind's platform lets enterprises and cloud providers
deliver secure, fully managed SLURM clusters as elastic, multi-tenant HPC services with lifecycle automation and governance using the NorthWind Platform
. It pairs Slurm scheduling with Kubernetes to give
dynamic GPU allocation, and multi-tenant security... secure workload separation with per-tenant tracking
. Combined with Project Slinky, it enables
secure, multi-tenant, self-service access to Slurm-based HPC environments on shared Kubernetes clusters
. (northwind.co)

2. ClearML — offers
built-in secure dynamic multi-tenancy with billing included, isolating teams, projects, and business units while they share the same physical infrastructure, with per-tenant resource limits and usage tracking
across Slurm and Kubernetes backends. (clear.ml)

3. OpenNebula (OneSlurm) — runs Slurm inside OpenNebula-managed VMs, providing
isolation, multi-tenancy, image-based provisioning, lifecycle management, and elastic scaling, while GPU access is provided through PCI passthrough
, aimed at AI Factories offering Slurm training environments. (opennebula.io)

4. Nebius (Managed Soperator) — an in-house Kubernetes operator that lets Nebius
quickly deploy thousand-GPU clusters for clients, giving a ready-to-work Slurm training cluster literally in minutes
. (nebius.com)

5. Together AI — designs multi-tenant GPU clusters where
each team gets a fully isolated virtual environment: dedicated GPU nodes, dedicated storage PVCs, and their choice of orchestration layer — Kubernetes, Slurm, or other configurations
. (together.ai)

Other notable managed-Slurm offerings include Crusoe (currently single-tenant per
Crusoe Managed Slurm on CMK is intentionally focused: single-tenant GPU training clusters
), Lambda (managed/unmanaged Slurm on HGX B200/H100), Runpod, and BUZZ HPC — worth evaluating if isolation requirements are less strict.