NVIDIA H100 SXM - 80GB usable
Specifications
| Brand | NVIDIA |
|---|---|
| Model | H100 SXM |
| Usable VRAM | 80GB |
| Architecture | Hopper |
| CUDA / Stream Processors | 16,896 |
| Memory Bandwidth | 3350 GB/s |
| TDP | 700W |
| FP32 TFLOPS | 67 |
Current Offers
No fresh used offer today · recently $20,000 on eBay
GPUDojo is reader-supported. When you buy through links on our site, we may earn an affiliate commission.
Lowest tracked offers by retailer
US · USD · Weekly since Sep 2026 · Shipping may be extra
- eBay (no current offer)
About these prices
Each line is the lowest offer we tracked from that retailer at weekly snapshots, ending at today's offer. The condition can vary between snapshots (new, used, or unrecorded for older ones). eBay prices exclude shipping. Tax is excluded. Amazon shows a current price only.
Weekly offers table
| Date | eBay |
|---|---|
| Oct 23, 2026 | $19,999.99 |
| Oct 9, 2026 | $19,999.99 |
| Oct 2, 2026 | $23,999.00 |
| Sep 25, 2026 | $37,998.99 |
| Sep 18, 2026 | $17,549.99 |
| Sep 11, 2026 | $17,549.99 |
For AI / LLM Use
Best for running 70B+ models on a single GPU. Very fast generation speed. Datacenter card with no display output, may need aftermarket cooling.
What Models Can It Run?
- 70B full precision
- 70B Q4_K_M, 30B full precision
- 30B Q6_K, 70B Q2_K
- 30B Q4_K_M, 14B full precision, 70B Q2 (tight)
- 14B Q6_K, 30B Q3_K (tight)
- 14B Q4_K_M, 7B full precision
- 7B Q6_K, 14B Q3_K (tight)
- 7B Q4_K_M only
Estimated Performance
Generation: ~251 tokens/sec
Prefill: ~1196 tokens/sec
Recommended Quantisations
- Q6_K or Q8 for best quality
- Q4_K_M for larger models
- Full precision for 30B and below
Pros & Cons
Pros
- 80GB usable VRAM: handles large models
- High memory bandwidth for fast generation
- Hopper architecture: broad CUDA software support
Cons
- 700W TDP: high power draw
- No display output: headless only
- May need aftermarket cooling solution
Community Verdict
No community reviews yet for the H100 SXM. Know a good review? Let us know.