Price check
Compare the lowest quote with the number of active providers and recent price history before selecting a platform.
NVIDIA H100 SXM5 80GB
80 GB VRAM989.5 TFLOPS FP16HopperNVIDIA
Compare H100 SXM5 hourly prices and availability across providers before renting.
Compare H100 rental providers →Last quote collected:
Starting from
$1.6000
per GPU / hour
Genesis Cloud · 2026-10-26
Confirm availability
Price Comparison Across Platforms
Latest collected quote per provider. Valid records from the last 48 hours appear first, then sorted by price. Check collection dates and confirm current rates and inventory with the provider.
Some quotes are older than 48 hours or have no collection date and are marked as historical references. Availability also comes from those records; confirm with the provider before renting.
| Platform | Price/Hour (USD) | Price/Hour (CNY) | Availability | Quote collected | Action |
|---|---|---|---|---|---|
Genesis CloudLowest recent quote | $1.6000 | ¥11.60 | Confirm availability | Check availability → | |
Shadeform | $1.6600 | ¥12.04 | Confirm availability | Check availability → | |
Latitude.sh | $1.6800 | ¥12.18 | Confirm availability | Check availability → | |
Gcore | $1.7900 | ¥12.98 | Confirm availability | Check availability → | |
Oblivus | $1.9800 | ¥14.36 | Confirm availability | Check availability → | |
Voltage Park | $1.9900 | ¥14.43 | Confirm availability | Check availability → | |
TensorDock | $2.2500 | ¥16.31 | Confirm availability | Check availability → | |
Crusoe Cloud | $2.9900 | ¥21.68 | Confirm availability | Check availability → | |
Paperspace | $3.0900 | ¥22.40 | Confirm availability | Check availability → | |
Hyperstack | $3.2000 | ¥23.20 | Confirm availability | Check availability → | |
Lambda Labs | $3.2900 | ¥23.85 | Confirm availability | Check availability → | |
Jarvis Labs | $3.4900 | ¥25.30 | Confirm availability | Check availability → | |
Daytona | $3.9500 | ¥28.64 | Confirm availability | Check availability → | |
DataCrunch | $4.6060 | ¥33.39 | Confirm availability | Check availability → | |
Replicate | $5.4900 | ¥39.80 | Confirm availability | Check availability → | |
vast.ai 1-GPU node · prices per GPU $1.33/node/hour Unverified host United Kingdom, GB | $1.3348 | ¥9.68 | Historical quote; recheck stock | Historical quote; needs checking | Check availability → |
Scaleway | $3.6600 | ¥26.54 | Confirm availability | Not recorded Quote date unavailable; verify before renting | Check availability → |
Modal | $3.9500 | ¥28.64 | Historical quote; recheck stock | Historical quote; needs checking | Check availability → |
7-Day Price History
Latest available daily quote per platform over the past 7 days.
H100 rental decision
H100 has broad provider coverage, so the best decision usually comes from comparing live price, availability, region, and platform terms for the exact SXM5 80GB configuration.
Workload fit
Evaluate H100 for training or high-throughput inference only after confirming memory and interconnect requirements.
Compare next
Compare H100 with H20 for 96GB inference capacity and B200 for newer high-memory infrastructure.
FAQ
How much does NVIDIA H100 SXM5 80GB rental cost?
NVIDIA H100 SXM5 80GB is quoted from $1.6000/GPU/hour on Genesis Cloud. Prices can change with inventory; check the collection date and provider billing terms.
What is NVIDIA H100 SXM5 80GB best for?
NVIDIA H100 SXM5 80GB has 80GB VRAM, making it ideal for LLM inference, AI training, and other high-performance computing workloads.
Which platform has the cheapest NVIDIA H100 SXM5 80GB?
Among currently tracked quotes, Genesis Cloud has the lowest NVIDIA H100 SXM5 80GB quote at $1.6000/GPU/hour. Check the collection date, availability and provider billing terms.
Should I rent an H100 or an H20?
The H100 is commonly evaluated for high-throughput AI workloads, while H20 offers 96GB VRAM. Compare workload fit, availability, and live hourly cost.
What makes one H100 rental offer better than another?
For the same SXM5 80GB identity, compare hourly price, availability, region, node configuration, networking, and platform terms.
Continue comparing