Back to all GPUs

NVIDIA H100 80GB SXM Cloud GPU Specs

HexGrid Cloud rents the NVIDIA H100 80GB SXM at $2.99 per GPU hour, billed per minute — about 6% below the $3.19/hr average across 1 other clouds. The H100 SXM carries 80GB of HBM3 memory on the Hopper architecture.

80GB VRAMHopperDeploys in ~1 minSecure Cloud

HexGrid Cloud price

$2.99/hr

Billed per minute

VRAM

80GB

HBM3

FP32

67 TFLOPS

Compute performance

Market average

$3.19/hr

You save 6%

Running the H100 SXM on HexGrid Cloud

What you get when you deploy with us, beyond the hourly rate.

Live in under a minute

Pick a region, hit deploy, and SSH in. No quota requests, no sales call.

Per-minute billing

You pay for the minutes you use. Stop the instance and billing stops with it.

Ready for training day one

CUDA, PyTorch and the usual drivers are preinstalled, or bring your own image.

No egress charges

Move checkpoints and datasets out without a surprise bandwidth bill.

H100 SXM pricing across the market

Published on-demand rates from other clouds, so you can see where our H100 SXM price sits.

Market average $3.19/hr across 1 cloud

6% below market average
  • HexGrid CloudYou're here$2.99/hr
  • RunPod$3.19/hr

Other providers' rates are their published on-demand list prices. They exclude committed-use discounts and can change at any time. Shown for reference only.

About the NVIDIA H100 80GB SXM

What the H100 SXM is built for, and where it falls short.

The NVIDIA H100 SXM is the high-performance SXM5 version of NVIDIA's Hopper accelerator. It combines 80GB HBM3, 3.35 TB/s memory bandwidth, fourth-generation Tensor Cores, FP8 Transformer Engine acceleration, MIG, and 900 GB/s NVLink.

Best suited for

The H100 SXM provides greater memory bandwidth, power headroom, and NVLink bandwidth than H100 PCIe and is optimized for tightly coupled multi-GPU AI systems.

  • Large-scale LLM training
  • LLM inference
  • Generative AI
  • Distributed training
  • HPC
  • Scientific computing

Strengths

  • 3.35 TB/s memory bandwidth
  • 900 GB/s NVLink
  • FP8 Transformer Engine
  • Excellent distributed training performance
  • MIG support

Limitations

  • 80GB memory capacity
  • High 700W power envelope
  • Requires SXM/HGX infrastructure

NVIDIA H100 80GB SXM specifications

Full technical specifications for the NVIDIA H100 80GB SXM.

VendorNVIDIAArchitectureHopperChipGH100CategoryDatacenter GPUReleased2022Form factorSXM5InterfaceSXM5

Memory

Memory
80 GB
Type
HBM3
Bandwidth
3,350 GB/s
Bus width
5120-bit
ECC
Yes

Compute

CUDA cores
16,896
Tensor cores
528
Tensor generation
Gen 4

AI & compute performance

FP64
34 TFLOPS
FP32
67 TFLOPS
TF32 Tensor
494.5 TFLOPS / 989 TFLOPS sparse
BF16 Tensor
989.5 TFLOPS / 1,979 TFLOPS sparse
FP16 Tensor
989.5 TFLOPS / 1,979 TFLOPS sparse
FP8 Tensor
1,979 TFLOPS / 3,958 TFLOPS sparse
INT8 Tensor
1,979 TOPS / 3,958 TOPS sparse

Power & physical

TDP
700 W
Form factor
SXM5
Cooling
Server-cooled

Interconnect

Interface
SXM5
NVLink
Yes
NVLink bandwidth
900 GB/s
Infinity Fabric
No

Virtualization

MIG
Yes
vGPU
Yes
SR-IOV
Yes

Media engines

NVENC engines
0
NVDEC engines
7
AV1 encode
No
AV1 decode
Yes

Software

CUDA
Yes
ROCm
No

Precision support

FP64
Yes
FP32
Yes
TF32
Yes
BF16
Yes
FP16
Yes
FP8
Yes
INT8
Yes

H100 SXM frequently asked questions

Common questions about NVIDIA H100 80GB SXM pricing, specifications and workloads on HexGrid Cloud.

How much VRAM does H100 SXM have?

The NVIDIA H100 SXM has 80GB of HBM3 memory.

What is the memory bandwidth of H100 SXM?

The H100 SXM provides 3.35 TB/s of GPU memory bandwidth.

How fast is H100 NVLink?

NVIDIA specifies 900 GB/s of NVLink bandwidth per H100 SXM GPU.

Is H100 SXM better than H100 PCIe for training?

H100 SXM generally provides higher compute throughput, memory bandwidth, and GPU-to-GPU interconnect bandwidth, making it better suited to large multi-GPU training.

How much does H100 SXM cost per hour?

Hourly pricing varies based on provider and cluster configuration. HexGrid compares available H100 SXM cloud rates.

Other GPUs on HexGrid Cloud

If the H100 SXM isn't the right fit, these are also available to deploy.

Browse every GPU we rent

Start training on the H100 SXM today

$2.99 per GPU hour, billed per minute, running in about a minute.

Specification sources