NVIDIA HGX B200 8-GPU Server

Name: NVIDIA HGX B200 8-GPU Server
Brand: NVIDIA
SKU: HGX B200

PLATFORMHGX B200

NVIDIA HGX B200 is an eight-GPU Blackwell platform with 1.4 TB of HBM3e and NVLink 5 at 1.8 TB/s per GPU. It keeps strong FP64 performance (296 TFLOPS per 8-GPU board), so it suits mixed AI and HPC clusters.

Source to orderConfirmed per quote

Price on written offer

Send quantity and destination. We confirm supply, lead time and a fixed price in writing.

Units required

Request a quote

Call sales +971 58 561 7897 Email this item

Exact part number HGX B200. No edition substitution without your written approval.
Stated per offer: manufacturer, OEM or supplier backed.
Lead time confirmed in your written offer. Agreed per order (Incoterms 2020).
30% deposit to secure the allocation, balance before dispatch. Bank transfer against pro forma invoice, or card payment link for approved orders. Card details are entered on Network International's secure payment page, never on our servers.
Lead-time orders are secured with a 30% deposit, balance before dispatch. Written offers are valid for the period stated, usually 48 to 72 hours.

Offer terms

Availability: Source to order
Unit price: Quoted per order
Minimum order: 1 unit
Ships from: Confirmed in your written offer
Dispatch: Lead time confirmed in your written offer
Incoterm: Agreed per order (Incoterms 2020)
Tax and duties: Quoted excluding UAE VAT, duties and import taxes unless stated
Payment: 30% deposit to secure the allocation, balance before dispatch; Bank transfer against pro forma invoice, or card payment link for approved orders
Condition: New, factory sealed

Overview

HGX B200 delivers 144 PFLOPS FP4 with sparsity (72 PFLOPS dense) and 72 PFLOPS FP8 across eight Blackwell GPUs with 180 GB of HBM3e each. Unlike Blackwell Ultra, it retains high double-precision throughput: 296 TFLOPS FP64 per board.

It ships inside OEM servers. We quote the exact OEM SKU, configuration and lead time for your site.

Typical uses

Training
Inference
HPC and simulation
Data analytics

Specifications

Platform
GPUs	8x NVIDIA Blackwell SXM
MIG	Up to 7 instances per GPU

Memory
GPU memory	1.4 TB HBM3e (180 GB per GPU)
Memory bandwidth	62 TB/s total (7.7 TB/s per GPU)

Performance
FP4 Tensor	144 PFLOPS sparse, 72 PFLOPS dense
FP8 and FP6 Tensor	72 PFLOPS (with sparsity)
FP64 and FP64 Tensor	296 TFLOPS

Interconnect
NVLink	5th generation, 1.8 TB/s per GPU, 14.4 TB/s total
Host interface	PCIe Gen5, 128 GB/s

Power and thermal
GPU power	Configurable up to 1,000 W per GPU
Cooling	Air-cooled and liquid-cooled systems available by OEM

Source: manufacturer documentation, checked 16 October 2026. Tensor figures follow NVIDIA's stated basis; sparse figures are marked.

NVIDIA HGX platform NVIDIA Blackwell datasheet (PDF)

Delivery and warranty

Delivery. Lead time confirmed in your written offer. Agreed per order (Incoterms 2020). Shipments are insured to declared value and tracked. Delivery to every destination is subject to export and sanctions screening; see export compliance.

Warranty. Stated per offer: manufacturer, OEM or supplier backed. Keep serial numbers with your invoice; our warranty and RMA process explains how claims are handled.

Cancellation. Written offers state their validity and cancellation terms. See returns and cancellations.

Questions about the HGX B200

B200 or B300?

B300 (Blackwell Ultra) has more memory per GPU (270 GB against 180 GB) and higher dense FP4 throughput, aimed at reasoning inference. B200 keeps far more FP64 performance, which matters for HPC and scientific workloads.

How quickly can you confirm availability?

Listed availability shows when supply was last verified. For other quantities or destinations, send a request and we confirm supply and price in writing.