HGX B300 pricing on Hyperstack
Every Hyperstack HGX B300 offering we track, normalized to a per-GPU hourly rate and ranked against 34 providers with current HGX B300 pricing.
The cheapest HGX B300 we currently track is $3.49/hr on Compute Cheap.
Hyperstack HGX B300 rates
Hourly prices per GPU. Multi-GPU rows show the per-GPU rate, not the instance total.
| Price / GPU-hr | Type | GPUs |
|---|---|---|
| $7.40 | On-demand | 1 |
HGX B300 specifications
- VRAM
- 288GB
- Architecture
- Blackwell
- Memory Bandwidth
- 8000 GB/s
- TDP
- 1100W
HGX B300 on other providers
- Compute Cheap$3.49/hr
- Aircloud$3.50/hr
- Daytona$4.08/hr
- Avian Compute$4.23/hr
- GPU Outlet$4.30/hr
- Bentaus$4.35/hr
- CoreWeave$4.48/hr
- TheAI Cloud$4.50/hr
- Impossible Cloud$4.53/hr
- Together AI$4.99/hr
- Civo$5.49/hr
- Seeweb$5.60/hr
- Verda$5.73/hr
- UpCloud$5.84/hr
- Deep Infra$5.89/hr
- DigitalOcean$6.17/hr
- Massed Compute$6.60/hr
- Gcore$6.73/hr
- Runpod$6.94/hr
- Modal$7.10/hr
- Latitude.sh$8.00/hr
- Scaleway$8.41/hr
- OpenRelay$8.49/hr
- Spheron$8.80/hr
- Crusoe$8.96/hr
- gpu.ai$8.99/hr
- Nebius$9.50/hr
- Lium.io$10.00/hr
- fal.ai$12.99/hr
- Runcrate$13.00/hr
- Fluence Cloud$13.66/hr
- Oracle Cloud$15.00/hr
- Amazon AWS$17.80/hr
About Hyperstack
Hyperstack offers on-demand, reserved and spot NVIDIA GPU VMs (B300, B200, H200, H100, A100, L40 and more) billed per minute, plus AI Studio for token-billed inference and fine-tuning.
Frequently Asked Questions
How much does the HGX B300 cost on Hyperstack?
Hyperstack lists the HGX B300 from $7.40 per GPU-hour across 1 tracked configuration. Prices are collected daily — see the table above for the current rates.
Is Hyperstack the cheapest place to rent the HGX B300?
Hyperstack ranks #21 of 34 providers we track with current HGX B300 pricing. The cheapest right now is Compute Cheap. Price is only one factor — region availability, network, and storage costs differ between providers.
Which HGX B300 configurations does Hyperstack offer?
We currently track 1 HGX B300 offering from Hyperstack, covering the GPU counts, pricing types, and regions listed in the table above. Every price is normalized to a per-GPU hourly rate so it can be compared across providers.