Europe
Availability on request
Price on requestConfiguration and region to be confirmed
GLOBALATLAS AI · B300
NVIDIA B300 · 288 GB
HGX B300 has 288 GB per GPU. Interconnect, GPU count and rack configuration are separate deployment requirements.
Blackwell Ultra288GB VRAM
- Memory technology
- HBM3e
- Workloads
- Large model inference and multi-GPU training
Memory capacity for larger working sets
- 01
288 GB HBM3e reference configuration
- 02
Model placement and communication between GPUs
- 03
Dataset delivery and checkpoint storage
Choose the complete HGX node topology alongside the GPU. A per-GPU memory figure does not specify a multi-GPU server.
Regional configurations
USA
Availability on request
Price on requestConfiguration and region to be confirmed
Asia
Availability on request
Price on requestConfiguration and region to be confirmed
Choosing B300
Does 288 GB describe the whole server?
No. It is the reference memory per GPU. The number of accelerators, system RAM, disks and network are specified separately for the HGX node.
Technical references
Specification 1 →YOUR CONFIGURATION
Request an infrastructure quote
Choose resources and operating time. We will confirm a price for your configuration.
Each resource has its own listed configuration. Multiple GPU servers do not form an NVLink cluster. Monthly rates cover the full billing period; storage and traffic outside the configuration are not included.