Flagship
Back to all servers
GPU 80 GB (H100) Flagship
The flagship. H100 80 GB Hopper. Best single-GPU performance available.
Commitment term
$50920.08
Billed once · 24-month commit · $2121.67/mo
Save ~24% vs 1-month
99.9% uptime SLA Provisions in < 15 min
Specifications
CPU
36 cores Dual E5-2697v4
RAM
256 GB
Storage
240 GB SSD + 2 TB NVMe + 8 TB SATA
VRAM
80 GB
What's included
- Hopper architecture
- Transformer Engine
- 3 TB/s HBM3
Best for
Train your own modelReal-time 70B inferenceResearch workloads
All pricing tiers
| Term | Per-month | Total commit |
|---|---|---|
| 1 month | $2791.67 /mo | $2791.67 |
| 3 months | $2707.92 /mo | $8123.76 |
| 12 months | $2326.39 /mo | $27916.68 |
| 24 months | $2121.67 /mo | $50920.08 |
You might also consider
4× A100 80 GB (320 GB total)
Training-cluster grade. Four A100 80GBs, 320 GB combined VRAM. Train your own foundation model.
$2449.90 /mo
GPU 80 GB (A100)
80 GB HBM2e — fits Llama 70B unquantized, all media warm together.
$2191.88 /mo
4× RTX A6000 (192 GB total)
192 GB VRAM. 512 GB system RAM. Heavy enterprise workload tier.
$1546.83 /mo