NVIDIA
H100 PCIe
Suitable for AI workflows such as AI chatbots, recommendation engines, and vision AI.
ArchitectureNVIDIA Hopper™
Cheapest Rent$1.99/hr
Hardware Intelligence Dossier
FP32
51 teraFLOPS
FP64
26 teraFLOPS
Decoders
7 NVDEC, 7 JPEG
GPU memory
80GB
Form factor
PCIe
Interconnect
NVLink: 600GB/s PCIe Gen5: 128GB/s
Server options
Partner and NVIDIA Certified Systems with 1–8 GPUs
FP8 Tensor Core
3,026 teraFLOPS
FP16 Tensor Core
1,513 teraFLOPS
FP64 Tensor Core
51 teraFLOPS
INT8 Tensor Core
3,026 TOPS
TF32 Tensor Core
756 teraFLOPS
Multi-instance GPUs
Up to 7 MIGs @ 10GB each
BFLOAT16 Tensor Core
1,513 teraFLOPS
GPU memory bandwidth
2TB/s
Max thermal design power (TDP)
300-350W (configurable)
Workload Suitability
The high-performance accelerator is engineered for mass-scale AI applications. From large language model (LLM) fine-tuning to high-throughput inference serving, this variant provides extreme memory bandwidth.
Asset Insights
"Preliminary specifications. May be subject to change."
Last Intel Refresh:February 2026
Verified Market Rates
Pricing Matrix
| Provider | Stock Status | Market Price | Network Link |
|---|---|---|---|
Civo Verified Node1d ago | High Availability | $1.99per hour | Inspect Listing |
RunPod Verified Node2d ago | High Availability | $2.39per hour | Inspect Listing |
RunPod Verified Node1d ago | High Availability | $2.39per hour | Inspect Listing |