B300
NVIDIA · Blackwell Ultra
288 GBHBM3e
8 TB/sBandwidth
4.5 PFFP8 dense
$4,899 /GPU/mo$6,552 · −25%
NVIDIA · Hopper · SXM · 8-GPU HGX
Rent the NVIDIA H100 SXM by the month.
The NVIDIA H100 SXM is the proven Hopper workhorse: 80 GB of HBM3, 1.98 PF of dense FP8 and NVLink 4 at 900 GB/s. It is the lowest-cost way to get an 8-GPU NVLink node for training, research and CUDA development.
80 GBHBM3 per GPU
3.35 TB/sMemory bandwidth
1.98 PFFP8 dense
NVLink 4Interconnect
Why the H100
What the H100 is best at
- 1.98 PF FP8 dense. The reference GPU most training recipes and benchmarks are tuned for.
- NVLink 4, 900 GB/s. Full 8-GPU NVLink mesh for data and tensor parallelism.
- Huge ecosystem. Every framework, kernel and tutorial supports it first.
- Budget pre-training and fine-tuning
- Research and ablations
- CUDA development and CI
Stock is low: when fewer than 8 GPUs are left, the configurator caps the server size accordingly.
Pricing
H100 monthly price by server size
One flat price per GPU, the same from 1 to 8 GPUs. The market median is the public on-demand list price of specialist GPU clouds × 730 hours.
| Server | CoreGPU / month | Market median / month | You save / month | You save / year | ≈ per GPU-hour |
|---|---|---|---|---|---|
| 1× H100 | $1,999 | $717 | $8,604 | $2.74 | |
| 2× H100 | $3,998 | $1,434 | $17,208 | $2.74 | |
| 4× H100 | $7,996 | $2,868 | $34,416 | $2.74 | |
| 8× H100 | $15,992 | $5,736 | $68,832 | $2.74 |
Prices in USD, excluding taxes. Market median checked Nov 4, 2026 across specialist GPU clouds; hyperscalers excluded. All GPU prices
Compare
H100 vs H200
Same compute as H200 with 80 vs 141 GB. Choose H100 when your model fits in 80 GB and price matters most.
| H100 | H200 | |
|---|---|---|
| Architecture | Hopper | Hopper |
| GPU memory | 80 GB HBM3 | 141 GB HBM3e |
| Memory bandwidth | 3.35 TB/s | 4.8 TB/s |
| FP8 dense | 1.98 PF | 1.98 PF |
| FP4 dense | n/a | n/a |
| Interconnect | NVLink 4 · 900 GB/s | NVLink 4 · 900 GB/s |
| Power (TDP) | 700 W | 700 W |
| Price per GPU / month | $1,999 | $2,699 |
| Price per GB of memory | $24.99 | $19.14 |
| Price per FP8 PFLOPS | $1,010 | $1,363 |
| Availability | Only 3 left | 34 available |
Specifications
NVIDIA H100 SXM technical specs
Per-GPU figures from the vendor datasheet. Tensor throughput is dense (without sparsity).
| Model | NVIDIA H100 SXM |
|---|---|
| Architecture | Hopper |
| Form factor | SXM · 8-GPU HGX |
| GPU memory | 80 GB HBM3 |
| Memory bandwidth | 3.35 TB/s |
| FP8 Tensor (dense) | 1.98 PF |
| FP4 Tensor (dense) | Not supported |
| Interconnect | NVLink 4 · 900 GB/s |
| Power (TDP) | 700 W |
| GPUs per server | 1 to 8 |
| Availability | Only 3 left |
FAQ
H100 rental: common questions
Another question? Our engineers answer within one business day.
Contact usHow much does it cost to rent an NVIDIA H100 SXM per month?
$1,999 per GPU per month at CoreGPU, about $2.74 per GPU-hour, versus a market median of $2,716 ($3.72/hour × 730 h). A full 8-GPU server is $15,992 per month, taxes excluded.
Is the H100 available right now?
Yes: only 3 left in the current allocation of 160 GPUs. In-stock servers are delivered with root SSH access in under 60 seconds after payment.
How many H100 GPUs can I rent?
From 1 to 8 GPUs per server, at the same price per GPU. Servers with 8 GPUs come with the full NVLink 4 mesh. Need more? Rent several servers or ask for a private allocation.
H100 or H200?
Same compute as H200 with 80 vs 141 GB. Choose H100 when your model fits in 80 GB and price matters most. See the full H100 vs H200 comparison.
What do I get with a H100 server?
Dedicated GPUs (never shared or sliced), root SSH access, your choice of a clean OS or a ready image with CUDA, PyTorch and vLLM, local NVMe scratch, and month-to-month billing with no setup fee.
Alternatives
Other GPUs in stock now
Deploy your H100 server in under 60 seconds
Pick 1 to 8 GPUs in the configurator above. Month to month, cancel anytime.