Worldwide supply, subject to product and destination approval +971 58 561 7897 sales@lineagetech.co
LINEAGE TECH

NVIDIA, HGX 8-GPU, new

NVIDIA HGX B300 8-GPU Server

PLATFORMHGX B300

NVIDIA HGX B300 is an eight-GPU Blackwell Ultra platform with 2.1 TB of HBM3e and fifth-generation NVLink at 1.8 TB/s per GPU, sold inside servers from the major OEMs. Lineage Tech quotes complete HGX B300 servers to your configuration.

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
  • Exact part number HGX B300. 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 B300 joins eight Blackwell Ultra SXM GPUs through an NVLink 5 switch, giving 14.4 TB/s of total GPU-to-GPU bandwidth. NVIDIA rates the platform at 144 PFLOPS FP4 with sparsity (108 PFLOPS dense) and 72 PFLOPS FP8, with 62 TB/s of aggregate memory bandwidth.

NVIDIA does not sell the baseboard on its own. It ships inside OEM servers that differ in CPU, memory, storage, networking, power and cooling. Send us your preferred OEM and configuration, or your workload and site limits, and we return a written offer with the exact manufacturer SKU and lead time.

Typical uses

  • Reasoning model inference
  • Large-scale training
  • AI factory building block

Specifications

Platform
GPUs8x NVIDIA Blackwell Ultra SXM
MIGUp to 7 instances per GPU
Memory
GPU memory2.1 TB HBM3e (270 GB per GPU)
Memory bandwidth62 TB/s total (7.7 TB/s per GPU)
Performance
FP4 Tensor144 PFLOPS sparse, 108 PFLOPS dense
FP8 and FP6 Tensor72 PFLOPS (with sparsity)
FP16 and BF16 Tensor36 PFLOPS (with sparsity)
FP32600 TFLOPS
Interconnect
NVLink5th generation, 1.8 TB/s per GPU, 14.4 TB/s total
Host interfacePCIe Gen6, 256 GB/s per GPU
Networking1.6 TB/s
Power and thermal
GPU powerConfigurable up to 1,100 W per GPU
CoolingAir-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.

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 B300

Can I buy an HGX B300 baseboard on its own?
No. NVIDIA supplies the baseboard to server manufacturers only. We quote complete servers from OEMs such as Supermicro, Dell, HPE, Lenovo, ASUS and Gigabyte.
Air or liquid cooling?
Both exist, depending on the OEM model and GPU power setting. Tell us your rack power and cooling and we will match the system.
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.

Related guides

Alternatives

NVIDIA, DGX systemSource to order
NVIDIA DGX B300PLATFORMDGX B300
  • 8x Blackwell Ultra, 2.1 TB HBM3e
  • 144 PFLOPS FP4 sparse, 108 dense
  • Turnkey 10U node for SuperPOD and BasePOD
Confirmed per quote
Price on written offer
Request a quote
NVIDIA, HGX 8-GPUSource to order
NVIDIA HGX B200 8-GPU ServerPLATFORMHGX B200
  • 8x Blackwell, 1.4 TB HBM3e
  • 144 PFLOPS FP4 sparse, 72 dense
  • Strong FP64 for mixed AI and HPC
Confirmed per quote
Price on written offer
Request a quote
NVIDIA, Rack-scaleSource to order
NVIDIA GB300 NVL72PLATFORMGB300 NVL72
  • 72 Blackwell Ultra GPUs, 36 Grace CPUs
  • 1,440 PFLOPS FP4 sparse, 1,080 dense
  • One 72-GPU NVLink domain, liquid-cooled
Confirmed per quote
Price on written offer
Request a quote

Often bought with