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

Several vendors now offer "SLURM-as-a-Service" specifically built for multi-tenant AI/HPC environments. Here are the leading options:

NorthWind GPU PaaS + Project Slinky –
NorthWind-powered SLURM as a Service enables organizations to deliver fully managed, multi-tenant SLURM environments for high-performance computing (HPC) workloads.
It uses
traditional SLURM scheduling with Kubernetes orchestration, dynamic GPU allocation, and multi-tenant security, launching clusters instantly with secure workload separation with per-tenant tracking.
Under the hood,
Project Slinky and NorthWind's GPU PaaS combined provide enterprises and cloud providers with a transformative combination that enables secure, multi-tenant, self-service access to Slurm-based HPC environments on shared Kubernetes clusters.
(northwind.co)

OpenNebula (OneSlurm) –
OneSlurm helps administrators create, scale, monitor, and operate Slurm clusters from OpenNebula, making it easier for AI Factories and HPC centers to offer Slurm-based training environments as part of a broader AI service portfolio.
Its underlying architecture
runs Slurm appliances inside OpenNebula-managed virtual machines, providing isolation, multi-tenancy, image-based provisioning, lifecycle management, and elastic scaling, while GPU access is provided through PCI passthrough.
A related feature, Elastic Capacity Management, ensures
each tenant runs in its own isolated environment, within separate virtual clusters, virtual networks, and access controls, so independent teams share the same physical infrastructure without interfering with each other's workloads.
(opennebula.io)

Nebius Managed Soperator –
Soperator, Nebius's in-house Kubernetes operator for Slurm, helps them quickly deploy thousand-GPU clusters for clients, simplifying deployment and cutting provisioning time from weeks to a couple of days.
(nebius.com)

Lambda Managed Slurm –
Lambda's Slurm workload scheduler offering includes both unmanaged and managed solutions for NVIDIA GB300 NVL72 and HGX B300 clusters, letting you choose unmanaged for full control, or managed to let Lambda handle administration.
(lambda.ai) Note this is oriented per-customer rather than explicitly multi-tenant.

Crusoe Managed Slurm – worth noting for contrast:
Crusoe Managed Slurm on CMK is intentionally focused on single-tenant GPU training clusters
, so it's not a multi-tenant option despite being a managed SLURM service. (crusoe.ai)

For true multi-tenant SLURM-as-a-Service, NorthWind/Slinky and OpenNebula/OneSlurm are currently the most explicitly designed for that use case, both layering SLURM atop Kubernetes/virtualization for tenant isolation, quotas, and self-service provisioning.