Data Center NVIDIA

NVIDIA H200 SXM

Hopper Architecture · 141GB HBM3e · SXM

VRAM
141GB
FP16
134.0
TDP
700W
Hardware Price
$38k
MSRP: $35k
Cloud from
$2.30/hr
7 providers
Cheapest at Fluidstack →

Quick Insights

Performance/Dollar
3.53 TFLOPS/$k
FP16 performance per $1000
VRAM/Dollar
3.7 GB/$k
VRAM per $1000
vs Data Center Average
-25% perf
FP16 TFLOPS comparison
Cloud Availability
7 providers
from $2.30/hr

Specifications

VRAM 141GB HBM3e
Memory Bandwidth 4.8 TB/s
FP16 TFLOPS 134.0
Tensor TFLOPS 2.0k
FP32 TFLOPS 67.0
TDP 700W
Form Factor SXM
Architecture Hopper
NVLink Yes (900GB/s)
Release Date 2024-03

Buy vs Rent Analysis

Buy Hardware
$38k
  • One-time cost, unlimited usage
  • Full control over hardware
  • Electricity & cooling costs extra
  • Depreciation over 2-3 years
Best if using >16522 hours total
Rent Cloud GPU
$2.30/hr
  • Pay only for what you use
  • No upfront investment
  • Scale up/down instantly
  • No maintenance required
Best for <16522 hours or variable usage
Breakeven Point
16,522
hours of usage

At $2.30/hr cloud pricing, buying the hardware pays off after 16,522 hours (~688 days or 22.9 months of 24/7 usage).

Usage Monthly Cloud Cost Months to Breakeven
100 hrs/month $230.00 166 months
200 hrs/month $460.00 83 months
500 hrs/month $1.1k 34 months

Cloud GPU Pricing

Rent NVIDIA H200 SXM from 7 cloud providers. Prices shown per GPU per hour.

Provider Type Instance GPUs On-Demand Per GPU Spot Availability
Fluidstack gpu-cloud fluidstack-h200-sxm 1x $2.30/hr $2.30/hr Cheapest - -
Genesis Cloud gpu-cloud genesis-h200-sxm 1x $2.95/hr $2.95/hr - -
Datacrunch (Verda) gpu-cloud datacrunch-h200-sxm 1x $2.99/hr $2.99/hr - -
RunPod gpu-cloud NVIDIA H200 1x $3.59/hr $3.59/hr $2.29/hr (-36%) -
Jarvis Labs gpu-cloud jarvislabs-h200-sxm 1x $3.99/hr $3.99/hr - -
CoreWeave gpu-cloud coreweave-h200-sxm 1x $6.31/hr $6.31/hr - -
Google Cloud Platform hyperscaler gcp-h200 1x $9.31/hr $9.31/hr $3.72/hr (-60%) -
Best Spot Deal: Google Cloud Platform offers spot pricing at $3.72/hr (60% off on-demand).

H200 vs Alternatives

Compare NVIDIA H200 SXM with similar GPUs from other brands.

GPU VRAM FP16 TFLOPS Bandwidth Hardware Price Cloud Price
H200 Current 141GB 134.0 4.8 TB/s $38k - -
MI300 AMD 128GB (-9%) 490.3 (+266%) 5.3 TB/s $15k (-61%) - Compare
AMD Instinct MI300A AMD 128GB (-9%) 980.0 (+631%) 5.3 TB/s - - Compare
AMD Instinct MI250X AMD 128GB (-9%) 383.0 (+186%) 3.3 TB/s - - Compare
AMD Instinct MI250 AMD 128GB (-9%) 362.0 (+170%) 3.3 TB/s - - Compare
MI300X AMD 192GB (+36%) 653.7 (+388%) 5.3 TB/s $18k (-53%) $1.99/hr Compare

Best Use Cases

No specific use case recommendations for NVIDIA H200 SXM yet.

Browse All Use Cases →

Compare H200

Popular Comparisons
Other NVIDIA GPUs
Alternatives
Add to Comparison Tool

Frequently Asked Questions about H200

The NVIDIA H200 SXM has a market price of approximately $38k (MSRP: $35k). Cloud rental starts at $2.30/hr. Prices may vary based on retailer, region, and availability.

Yes, the NVIDIA H200 SXM with 141GB VRAM is suitable for many AI/ML workloads. For large language models, you may need multiple GPUs or consider higher-VRAM options like A100 or H100.

The breakeven point is approximately 16,522 hours of usage. Buy if you'll use it more than this; rent for shorter projects or variable workloads. Cloud rental from Fluidstack starts at $2.30/hr.

With 141GB VRAM and 134.0 FP16 TFLOPS, the NVIDIA H200 SXM can run: Large language models (7B-13B), Stable Diffusion XL, video AI, and professional 3D rendering.

The NVIDIA H200 SXM offers 141GB VRAM and 134.0 FP16 performance at $38k. Compare with similar GPUs using our comparison tool above. Key factors: VRAM for model size, TFLOPS for speed, and price for budget.