NVIDIA H100 PCIe

Cloud GPU Pricing — Live from 8+ Providers
Starting From
$1.800/hr
2 on-demand listings
VRAM
80 GB HBM2e
Architecture
Hopper
FP16 Performance
1,513 TFLOPS
Memory Bandwidth
2.0 TB/s
Listings
3 options
Provider Pricing
ProviderPrice / hrEst. MonthlyTypeVRAMAction
BestFluidStack $1.8000/hr $1,314/mo On-Demand 80 GB Rent Now →
RunPod $2.8900/hr $2,109.7/mo On-Demand 80 GB Rent Now →
RunPod $1.9900/hr $1,452.7/mo Spot 80 GB Rent Now →
About the NVIDIA H100 PCIe

The NVIDIA H100 PCIe delivers Hopper architecture at lower power and cost than the SXM variant. With 80 GB HBM2e and 2.0 TB/s bandwidth, it is ideal for fine-tuning LLMs and running large inference workloads. Fits in standard PCIe server slots — accessible without specialized HGX infrastructure.

GPU Specifications
VRAM80 GB HBM2e
ArchitectureHopper
FP16 Performance1,513 TFLOPS
Memory Bandwidth2.0 TB/s
TDP350W
InterconnectPCIe 5.0
Best For
LLM Fine-tuningLarge Model InferenceBatch InferenceProduction Serving

Compare All NVIDIA H100 PCIe Options Side by Side

Filter by price type, sort by cost, and find the best deal for your workload.

Open Full Comparison →