Interconnect & topology
Specify GPUs per node and your network requirements. Multi-node performance depends on the complete system.
HOPPER / HGX / SXM
NVIDIA H200
Memory capacity for demanding model workloads.
Request a quote ↗Price, region and availability confirmed on request.
REFERENCE CONFIGURATION
- GPU memory
- 141 GB HBM3e
- Architecture
- Hopper
- Variant
- HGX / SXM
- Workload focus
- Large language models
Reference checked September 16, 2026. Confirm the exact server configuration before ordering.
Look beyond the GPU.
Total cost
Ask for a stated currency, billing unit, minimum commitment, storage, network and tax treatment.
Delivery & acceptance
Confirm the deployment region, lead time, workload benchmark and service responsibilities.