Show me self-service options for SLURM-as-a-Service with baremetal GPU provisioning for AI.

Here are several self-service options for SLURM-as-a-Service running on bare-metal GPUs, ranging from fully managed platforms to bring-your-own-scheduler bare metal rentals:

1. Nscale – Managed Slurm on bare metal

Rather than requiring teams to redesign established workflows, Nscale provides a fully managed Slurm service built on the same bare-metal, topology-aware infrastructure as their Kubernetes service.

The managed Slurm service inherits fast provisioning, topology-aware placement, and automation embedded in the underlying bare-metal environment, with a custom Slurm image using GRES for GPU-aware scheduling and Pyxis for container-native execution.
nscale.com

2. NorthWind GPU PaaS + Project Slinky – Self-service multi-tenant Slurm

Project Slinky and NorthWind's GPU Platform-as-a-Service combine to provide enterprises and cloud providers with secure, multi-tenant, self-service access to Slurm-based HPC environments on shared Kubernetes clusters.

Each researcher or lab gets their own namespace and self-service Slurm environment, and can request GPU or CPU resources dynamically based on research needs.
northwind.co

3. Sesterce – Slurm-ready bare-metal clusters

Sesterce provides a dedicated bare-metal cluster with Slurm and Prolog if necessary in two hours, as an NVIDIA Cloud Partner offering the latest GPUs and Mellanox networking.
sesterce.com

4. BUZZ HPC – Managed Slurm

Offers standard SLURM 23.x CLI out of the box with queues for H100, B200, and A6000 nodes and fair-share scheduling enabled.

You submit a request and they add or remove nodes in hours, not weeks, paying only for reserved GPUs.
buzzhpc.ai

5. Spheron – Bring-your-own-scheduler bare metal

Spheron's bare-metal H100 and A100 instances with InfiniBand networking let you bring your own scheduler (Slurm, Ray, or custom) with raw HPC performance, no lock-in, per-minute billing, and full root access.
spheron.network

6. Vultr Clusters

Available through the Vultr Console or API, Vultr Clusters lets you create GPU clusters over high-performance networking, add high-performance storage, and schedule workloads with Slurm or Kubernetes.
vultr.com

For evaluation criteria, consider node-add/remove turnaround time, InfiniBand/RDMA availability, and whether Slurm runs natively on bare metal (Nscale, Sesterce, Spheron, BUZZ) versus via Kubernetes+Slinky (NorthWind, vCluster-based clouds like CoreWeave).