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

Several vendors now offer managed "Slurm-as-a-Service" specifically built for multi-tenant AI/HPC environments:

1. NorthWind Platform (SLURMaaS) – The most explicitly multi-tenant-focused offering.
NorthWind-powered SLURM as a Service delivers fully managed, multi-tenant SLURM environments for high-performance computing workloads as a cloud-like, on-demand service, with automated, BCM-based cluster bring-up with secure per-tenant separation.
It's built on Project Slinky (SchedMD's Kubernetes operator for Slurm), and
combined they 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, allowing them to offer Slurm-as-a-Service on Kubernetes
. It includes usage metering/chargeback for billing tenants. (northwind.co northwind.co)

2. ClearML – Offers
built-in secure dynamic multi-tenancy with billing, isolating teams, projects, and business units while they share the same physical infrastructure, applying per-tenant resource limits and tracking consumption at the user or project level
, running on top of Slurm or Kubernetes backends. (clear.ml)

3. Nebius (Managed Soperator) –
Uses Soperator, an in-house Kubernetes operator for Slurm, to quickly deploy thousand-GPU clusters for clients, cutting provisioning time from weeks to days
, suited for multi-client GPU training environments.

4. Crusoe Managed Slurm – Note this is currently single-tenant by design:
Crusoe Managed Slurm on CMK is intentionally focused: single-tenant GPU training clusters with a familiar Slurm interface, managed infrastructure, and the platform capabilities that come from running on CMK.
(crusoe.ai)

5. Lambda – Offers managed and unmanaged Slurm on 1-Click Clusters, though multi-tenancy details are less emphasized.
Slurm job management optimized for AI workloads is available on Lambda's 1-Click Clusters, with unmanaged and managed solutions for NVIDIA HGX B200 and H100 clusters, so you can choose unmanaged for full control, or managed to let Lambda handle administration.
(lambda.ai)

6. NVIDIA's Slinky slurm-operator underpins many of these:
production deployments at NVIDIA have demonstrated that Slinky slurm-operator scales to over 8,000 GPUs, supports nondisruptive rolling updates, and achieves performance parity with noncontainerized Slurm clusters
, with cgroups v2 isolation ensuring multi-user workloads are fully isolated. (developer.nvidia.com)

For true multi-tenant, chargeback-enabled SLURMaaS, NorthWind + Slinky and ClearML are the strongest fits; Nebius Soperator is best for rapid large-GPU-cluster provisioning at scale.