GPU AI

Dedicated GPUs for your own models

RTX 4090, RTX 5090 and H100-class servers with CUDA, drivers and JupyterLab ready. SSH in and start working.

Prices in USD, billed in crypto. Save up to 50% on longer terms.

CUDA ready

NVIDIA drivers, CUDA and cuDNN preinstalled on Ubuntu images.

SSH + JupyterLab

Use the terminal or notebooks, whichever you prefer.

Whole card

Each GPU is passed through to you alone. No time-slicing.

Private inference

Your prompts and weights stay on your machine.

Plans by location

GPU stock is limited to selected locations.

GPU RTX 4090 IS-GPU-4090	1 × RTX 4090 · 24 GB GDDR6X · 16 vCPU · 64 GB RAM · 1 TB NVMe	$420/mo	Order
GPU RTX 5090 IS-GPU-5090	1 × RTX 5090 · 32 GB GDDR7 · 24 vCPU · 96 GB RAM · 2 TB NVMe	$690/mo	Order
GPU H100 IS-GPU-H100	1 × H100 80 GB (PCIe) · 32 vCPU · 192 GB RAM · 4 TB NVMe	$2,700/mo	Order

GPU RTX 4090 RO-GPU-4090	1 × RTX 4090 · 24 GB GDDR6X · 16 vCPU · 64 GB RAM · 1 TB NVMe	$420/mo	Order
GPU RTX 5090 RO-GPU-5090	1 × RTX 5090 · 32 GB GDDR7 · 24 vCPU · 96 GB RAM · 2 TB NVMe	$690/mo	Order
GPU H100 RO-GPU-H100	1 × H100 80 GB (PCIe) · 32 vCPU · 192 GB RAM · 4 TB NVMe	$2,700/mo	Order

GPU RTX 4090 CH-GPU-4090	1 × RTX 4090 · 24 GB GDDR6X · 16 vCPU · 64 GB RAM · 1 TB NVMe	$470.50/mo	Order
GPU RTX 5090 CH-GPU-5090	1 × RTX 5090 · 32 GB GDDR7 · 24 vCPU · 96 GB RAM · 2 TB NVMe	$773/mo	Order
GPU H100 CH-GPU-H100	1 × H100 80 GB (PCIe) · 32 vCPU · 192 GB RAM · 4 TB NVMe	$3,024/mo	Order

GPU RTX 4090 NL-GPU-4090	1 × RTX 4090 · 24 GB GDDR6X · 16 vCPU · 64 GB RAM · 1 TB NVMe	$420/mo	Order
GPU RTX 5090 NL-GPU-5090	1 × RTX 5090 · 32 GB GDDR7 · 24 vCPU · 96 GB RAM · 2 TB NVMe	$690/mo	Order
GPU H100 NL-GPU-H100	1 × H100 80 GB (PCIe) · 32 vCPU · 192 GB RAM · 4 TB NVMe	$2,700/mo	Order

Read about running open-weight models on your own hardware.