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" backed by bare-metal GPUs for AI workloads:

1. Runpod Clusters — Lets you
use Slurm to manage multi-node jobs with familiar sinfo, sbatch, and srun commands for distributed training and HPC workloads
, with
complete freedom to terminate your cluster when not in use, with no termination fees or minimum runtime requirements, perfect for intermittent workloads
. You can
spin up hundreds of GPUs with a single command
. → runpod.io

2. Spheron Network — Offers
bare-metal H100 and A100 instances with InfiniBand networking built for Slurm-style workloads, where you bring your own scheduler and get raw HPC performance without managed-Kubernetes overhead
, with
no lock-in, per-minute billing, full root access
. → spheron.network

3. Nscale Managed Slurm —
Built on the same virtual cluster architecture as NKS, inheriting fast provisioning, topology-aware placement, and automation embedded in the underlying bare-metal environment
, using
a custom Slurm image with GPU-aware scheduling enabled through Generic Resource Scheduling (GRES) and container-native execution via Pyxis
. → nscale.com

4. NorthWind GPU PaaS (with Slinky) — Provides
secure, multi-tenant, self-service access to Slurm-based HPC environments on shared Kubernetes clusters, allowing cloud providers and enterprise platform teams to offer Slurm-as-a-Service without compromising on performance
.
Researchers and ML teams can launch their own GPU-accelerated Slurm clusters in minutes—no manual provisioning, no tickets.
→ northwind.co

5. Sesterce Reserved Cloud —
Provides a dedicated bare-metal cluster with Slurm and Prolog in two hours, with the latest generation GPUs and Mellanox networking
. → sesterce.com

6. Buzz HPC Managed Slurm — Delivers
bare-metal GPU horsepower with zero scheduler upkeep, offering the shortest path from research idea to results with no data-center build-out required
. → buzzhpc.ai

For pure DIY control, Spheron/CoreWeave-style bare-metal rentals suit teams wanting full root access; for turnkey multi-tenant self-service, NorthWind/Nscale are stronger fits.