NovaPod LogoNovaPod
datacenterLive spot market active

NVIDIA H100 SXM

Industry-leading GPU for AI training with 80GB HBM3 and NVLink support.

Live Starting Ask
$2.60/ hour (billed per second)

Dynamic ask based on current network node availability.

Launch on Marketplace

Hardware Specifications

VRAM
80 GB
CUDA Cores
16,896
FP32 Compute
67 TFLOPS
Bandwidth
3,350 GB/s
Memory Type
HBM3
Architecture
Hopper

Supported Compute Modes on NovaPod

Docker GPU Pod

Ideal for running PyTorch, ComfyUI, or vLLM in a lightweight container. Boots in under 15 seconds with zero operating system overhead and zero idle fees when terminated.

  • 1-second streaming meter deduction
  • JupyterLab & VS Code web IDE access
Root KVM VPS

Full virtual machine with dedicated Ubuntu or Windows OS, sudo root SSH, and dedicated public IPv4. Pause anytime while preserving your NVMe system drive at just $0.00014/GB/hr.

  • Full root privileges & custom kernel modules
  • Open all ingress/egress ports directly

Non-Custodial Multi-Chain Crypto Settlement: Fund your dedicated personal address with USDT, USDC, SOL, ETH, or BNB. Zero KYC required.

Deposit Crypto →
← All GPU Pricing TablesBrowse All Live Asks →