Aquanode DocsCloud Features
GPU Types & Pricing
Compare GPU options, capabilities, and pricing across providers
Different workloads require different GPUs. Below you can compare available GPU models, their specifications, and pricing across providers.
GPU Catalog
| GPU Model | Memory | Ideal Use Case | Providers |
|---|---|---|---|
| NVIDIA H100 | 80 GB | Large LLM training | Akash, DataCrunch, Voltage Park |
| NVIDIA A100 | 40/80 GB | Training & inference | Akash, DataCrunch |
| RTX 6000 ADA | 48 GB | Inference, CV, smaller models | DataCrunch |
| NVIDIA B200 | 192 GB | Next-gen LLM workloads | DataCrunch |
Pricing Overview
| Provider | GPU Model | On-Demand Price/hr | Reserved Price | Notes |
|---|---|---|---|---|
| Akash | A100 80 GB | varies (auction) | N/A | Reverse auction model |
| DataCrunch | H100 80 GB | ~$2.5/hr | Monthly contract | Dynamic pricing available |
| Voltage Park | HGX H100 | $1.99/hr | Reserved 6+ months | Enterprise-grade clusters |
| Hot Aisle | MI300x | $1.95/hr | N/A | AMD GPU provider with on-demand pricing |
Pricing Models Explained
- Akash → Decentralized auction marketplace, often cheapest
- DataCrunch → Fixed + dynamic pricing, contract options
- Voltage Park → On-demand + long-term reservations for scaling