SwissGPU / Bare-metal compute
Rent a dedicated GPU server.
Hosted in Switzerland.
Serious compute. Swiss foundations.
Rent NVIDIA RTX 4090, RTX 5080 and RTX 5090 bare-metal servers for AI inference, image generation, rendering and development. Get the whole machine with Linux and SSH access; choose your software and keep control of your environment.
Rent a serverChoose your NVIDIA GPU
Compare GPU memory, server specifications and current hourly prices before signing in.
Hosted here.
Operated here.
Our servers, data, and operations are based in Switzerland. Talk directly with us about your requirements.
The whole machine.
For your workload.
Dedicated GPU, CPU, memory, and storage. Direct hardware access for AI, rendering, and development.
Clear rates.
Clear next steps.
Browse servers, check the billing details, and follow the documentation to get started.
Choose your next GPU server.
Compare on-demand pricing and see where fixed-term prepaid offers lower your compute rate.
Put your GPU to work
ComfyUI and image generation
Install your own image-generation tools, models and custom nodes on a dedicated server. Follow the setup guide and connect through an SSH tunnel.
ComfyUI setup guideLLM inference with Ollama
Choose a model that fits the GPU memory, run it on your server and access its API from your own tools. Software and model setup are yours to manage.
Ollama setup guidePyTorch and GPU development
Use a full Linux machine for experiments, development and workloads that fit one GPU. Check your CUDA and framework requirements before installing packages.
PyTorch setup guideAutomating your workload? Read offers and use the customer API.

