How much does an H100 SXM cost per hour to rent?

Runcrate H100 SXM rental is $1.35-1.65/hr on-demand (avg $1.50/hr). That's significantly below AWS ($4.10), GCP ($3.67), Azure ($3.96), and Oracle ($2.40). Reserved pricing drops the rate another 30%.

What's the difference between H100 SXM and H100 PCIe?

H100 SXM has 67% more memory bandwidth (3350 vs 2000 GB/s), full NVLink support for 8-GPU clusters, and 18% higher FP32 throughput. PCIe is cheaper and works in standard servers but caps at 4-GPU nodes without NVLink. SXM is the right choice for any multi-GPU workload.

Can I rent H100 SXM on demand without a contract?

Yes — Runcrate H100 SXM is fully on-demand with per-second billing across 8 regions. No contracts, no minimums, no commitments.

H100 vs A100 — when should I pick H100?

H100 has ~3x the FP16 throughput of A100 (120 vs 78 TFLOPS) and supports FP8. For modern transformer training, fine-tuning, or any inference workload, H100 is dramatically more cost-efficient. A100 only makes sense for legacy code that can't use FP8 or bf16.

How fast can I deploy an H100 SXM instance?

Most users deploy in under 60 seconds. Pick a region, choose a base image (vLLM, Comfy, Jupyter, custom), attach a volume, and SSH in. Per-second billing means there's zero risk to spinning up briefly.

runcrate

Contact Sales Console

NVIDIA · Hopper · 2022

NVIDIA H100 SXM.

Name: NVIDIA H100 SXM GPU Cloud Instance
Brand: NVIDIA
Price: 1.50 USD
Availability: InStock

The workhorse for production AI training and inference — the most-deployed Hopper GPU.

80 GB

HBM3

3.4 TB/s

Bandwidth

120

TFLOPS FP16

8× GPU

NVLink

Available now · 8 regions

Deploy in 60 seconds Reserved quote

Cloud rental

$1.50/hr

On-demand · per-second billing

$1.35–$1.65/hr range

Or $1.05/hr reserved (30% off)

Pricing across clouds

NVIDIA H100 SXM cloud rental price comparison.

Same GPU, cheapest-first. Prices reflect publicly listed hourly rates for NVIDIA H100 SXM on each provider. Runcrate is the lowest published rate.

RuncrateCheapest

$1.50/hr

RunPod

$1.99/hr

Oracle

$2.40/hr

Lambda

$2.49/hr

GCP

$3.67/hr

Azure

$3.96/hr

AWS

$4.10/hr

Workloads

What you'll actually use this for.

Real workloads sized for the NVIDIA H100 SXM, with concrete performance numbers. Click to deploy preconfigured.

Llama 70B Fine-tune

QLoRA in 12 hrs on 1 GPU

Deploy

Mistral 7B Inference

~480 tok/s in bf16

Deploy

Stable Diffusion XL

1024x1024 in 1.2s

Deploy

Multi-GPU Training

8x H100 SXM with NVLink

Deploy

One command. NVIDIA H100 SXM in 60 seconds.

Skip the dashboard if you don't need it. SDK, Python, or cURL — copy the snippet, paste your API key, ship.

import Runcrate from "@runcrate/sdk";

const rc = new Runcrate({ apiKey: "rc_live_••••••••••••••••" });

const instance = await rc.instances.create({
  gpu: "h100-sxm",
  region: "auto",
  image: "runcrate/vllm:latest",
});

console.log(`SSH: ssh root@${instance.host}`);