B300
NVIDIA · Blackwell Ultra
288 GBHBM3e
8 TB/sBandwidth
4.5 PFFP8 dense
$4,899 /GPU/mo$6,552 · −25%
NVIDIA · Blackwell · SXM · 8-GPU HGX
Rent the NVIDIA HGX B200 by the month.
The NVIDIA B200 is the mainstream Blackwell training GPU: 180 GB of HBM3e at 8 TB/s and 4.5 PF of dense FP8, with full FP64 support. It is the safe default for pre-training and large fine-tunes on the most mature Blackwell software stack.
180 GBHBM3e per GPU
8 TB/sMemory bandwidth
4.5 PFFP8 dense
9 PFFP4 dense
Why the B200
What the B200 is best at
- 4.5 PF FP8 dense. 2.3× an H100 on FP8 training throughput.
- 180 GB HBM3e at 8 TB/s. More than double the bandwidth of H100 for memory-bound kernels.
- Strong FP64. Unlike B300, keeps the double-precision throughput HPC codes need.
- Pre-training and continued pre-training
- Full fine-tuning of 30B–70B models
- Mixed AI and scientific computing
B200 allocations sell out fastest: join the waitlist to get the restock email before public release.
Pricing
B200 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× B200 | $4,449 | $1,530 | $18,360 | $6.09 | |
| 2× B200 | $8,898 | $3,060 | $36,720 | $6.09 | |
| 4× B200 | $17,796 | $6,120 | $73,440 | $6.09 | |
| 8× B200 | $35,592 | $12,240 | $146,880 | $6.09 |
Prices in USD, excluding taxes. Market median checked Nov 4, 2026 across specialist GPU clouds; hyperscalers excluded. All GPU prices
Compare
B200 vs H200
B200 is roughly 2.3× faster than H200 on FP8 with more memory. H200 remains cheaper per month when Hopper performance is enough.
| B200 | H200 | |
|---|---|---|
| Architecture | Blackwell | Hopper |
| GPU memory | 180 GB HBM3e | 141 GB HBM3e |
| Memory bandwidth | 8 TB/s | 4.8 TB/s |
| FP8 dense | 4.5 PF | 1.98 PF |
| FP4 dense | 9 PF | n/a |
| Interconnect | NVLink 5 · 1.8 TB/s | NVLink 4 · 900 GB/s |
| Power (TDP) | 1,000 W | 700 W |
| Price per GPU / month | $4,449 | $2,699 |
| Price per GB of memory | $24.72 | $19.14 |
| Price per FP8 PFLOPS | $989 | $1,363 |
| Availability | Back Nov 20 | 34 available |
Specifications
NVIDIA HGX B200 technical specs
Per-GPU figures from the vendor datasheet. Tensor throughput is dense (without sparsity).
| Model | NVIDIA HGX B200 |
|---|---|
| Architecture | Blackwell |
| Form factor | SXM · 8-GPU HGX |
| GPU memory | 180 GB HBM3e |
| Memory bandwidth | 8 TB/s |
| FP8 Tensor (dense) | 4.5 PF |
| FP4 Tensor (dense) | 9 PF |
| Interconnect | NVLink 5 · 1.8 TB/s |
| Power (TDP) | 1,000 W |
| GPUs per server | 1 to 8 |
| Availability | Sold out · next allocation Nov 20 |
FAQ
B200 rental: common questions
Another question? Our engineers answer within one business day.
Contact usHow much does it cost to rent an NVIDIA HGX B200 per month?
$4,449 per GPU per month at CoreGPU, about $6.09 per GPU-hour, versus a market median of $5,979 ($8.19/hour × 730 h). A full 8-GPU server is $35,592 per month, taxes excluded.
Is the B200 available right now?
Not right now. The current allocation is sold out and the next one opens on November 20, 2026. Join the waitlist and we email you before units go back on public sale.
How many B200 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 5 mesh. Need more? Rent several servers or ask for a private allocation.
B200 or H200?
B200 is roughly 2.3× faster than H200 on FP8 with more memory. H200 remains cheaper per month when Hopper performance is enough. See the full B200 vs H200 comparison.
What do I get with a B200 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
Get the B200 restock email first
The next allocation opens November 20. Waitlist members are emailed before public release.