About Hyperstack
Hyperstack is a cloud GPU provider for AI and machine learning teams. It offers on-demand virtual machines with NVIDIA GPUs, along with object storage, Kubernetes and a managed AI studio.
What it does
You pick a GPU model, launch a virtual machine and run training or inference jobs. The site lists several NVIDIA GPU types, from data center cards to workstation options. A spot VM option is available for flexible workloads. SDKs for Python and Go, plus a Terraform provider, let you script the setup of machines and storage instead of clicking through a console.
Key features
- NVIDIA GPU machines in several models
- Spot virtual machines for flexible jobs
- On-demand Kubernetes clusters
- Python and Go SDKs and a Terraform provider
Who it's for
AI teams that train or serve models and need GPU capacity without buying hardware. Simulation, visualization and high-performance computing teams can also use it for demanding jobs and remote virtual workstations.
Pricing and good to know
The site has a GPU pricing page, so check it for current rates before you start a large job. Spot machines usually cost less than standard ones, but they can be reclaimed, so plan checkpoints for long training runs. A support hub and a documentation section cover setup questions.
Highlights
- NVIDIA GPU virtual machines in several models
- Spot VMs for interruptible workloads
- On-demand Kubernetes for container jobs
- Python and Go SDKs and a Terraform provider
