Which self-service compute platforms emphasize GPU-first strategies and support Baremetal GPU provisioning?

Several platforms explicitly position themselves as "GPU-first" and offer self-service bare metal GPU provisioning:

1. Lambda (Lambda Bare Metal Instances) – A GPU cloud built specifically for AI workloads.
Launching a Lambda Bare Metal Instance looks nearly identical to launching a virtualized one, but instead of selecting a virtualized instance type, you select a bare metal type, giving direct, unmediated access to the CPU, memory, GPU, disk, and TPM.

Bare Metal Instances are coupled with self-service APIs covering the instance lifecycle, VPCs, storage volumes, and resource governance.
(lambda.ai)

2. vCluster Labs (vMetal + vCluster + vNode) – A GPU-focused infrastructure stack for neoclouds/AI factories.
vMetal handles the bare metal provisioning lifecycle—PXE boot, OS installation, network configuration—while vCluster Standalone provides the Kubernetes distribution running directly on that hardware
, and
vNode gives each workload its own secure runtime for container breakout protection at bare metal GPU performance with no hypervisor overhead.
(vcluster.com)

3. NorthWind (BMaaS – Bare Metal GPUs-as-a-Service) –
Deploys dedicated NVIDIA GPUs instantly with no virtualization, full control, and cloud-like flexibility
, supporting
NVIDIA H100, A100, L40S and more, up to 8 GPUs per server with NVLink and NVSwitch.
(northwind.co)

4. Oracle Cloud Infrastructure (OCI) Compute –
Provides industry-leading scalability and performance for bare metal and virtual machine (VM) instances powered by NVIDIA and AMD GPUs for mainstream graphics, AI inference, AI training, digital twins, and HPC.
(oracle.com)

5. Vultr (Vultr Clusters/Bare Metal) –
Delivers on-demand, self-service Vultr Bare Metal server clusters, provisioned through Vultr's console or API, providing scalable and customizable CPU and GPU compute solutions.
(vultr.com)

6. phoenixNAP Bare Metal Cloud –
Lets users quickly spin up workload-optimized dedicated GPU servers at different locations using API, CLI, IaC, or the platform's WebUI.
(phoenixnap.com)

Note that raw bare-metal hardware providers like Equinix Metal, DigitalOcean, and OVHcloud offer GPU bare metal but leave provisioning/orchestration largely to the customer, per vCluster's analysis.