Mfr Part #:

A100

Mfr Name: NVIDIA

The NVIDIA DGX A100 System is a universal system purpose-built for all AI infrastructure and workloads. It features eight NVIDIA A100 Tensor Core GPUs.

✓ Total Stock: 1,500 pcs

$ Price Range: $0.99

A100 Available from 2 distributors

✓ Authorised

✓ Non-Authorised

Distributor		Stock	Pricing	MOQ	PKG	Lead Time	Date Code	On Order Quantity	On Order Delivery
Xingangxin ✓ Non-Authorised Inventory information	1,000 pcs On Request	1000 pcs	On Request	1	On Request	On Request	On Request	On Request	On Request
PROSOFT ELECTRONICS ✓ Non-Authorised Inventory information	500 pcs On Request	500 pcs	On Request	1	On Request	On Request	22+	On Request	On Request

A100 Specifications Quick View

Most important parameters for selection – full specs in datasheet.

SELECT ALL

Category	New Products
Manufacturer Name	NVIDIA
Depth	897.1 mm
Form Factor	6U Rackmount
Height	264 mm
Input Voltage	200-240 V
Width	482.3 mm

View Similar Products (0)

A100 Description

Short technical summary and product information.

The NVIDIA DGX™ A100 System is engineered as a universal solution for all AI infrastructure and workloads, spanning analytics, training, and inference. This powerful system is built upon eight NVIDIA A100 Tensor Core GPUs, providing immense computational power.

Key hardware components include 8 NVIDIA A100 GPUs connected via third-generation NVLinks, offering a total of 640 GB of GPU memory. It utilizes six second-generation NVIDIA NVSwitch units for high-speed GPU-to-GPU communication with 600 GB/s bandwidth. The system is equipped with 2 TB of system memory and 30 TB of U.2 NVMe storage for data caching and OS.

Networking capabilities are robust, featuring up to 10 factory-configured NVIDIA ConnectX-6 or ConnectX-7 InfiniBand/200 Gb/s Ethernet ports. The system is housed in a 6U rackmount form factor with dimensions of 10.4” height, 19” width, and 35.3” depth, weighing up to 271.5 lbs. Power is supplied by six 3 kW power supplies configured for N+N redundancy, operating within a 200-240V AC input range.

A100 Quick Information

Product Type: Server Acceleration Card

Series: DGX A100

Canonical Name: NVIDIA DGX A100 System

Subcategory: DGX A100 System

A100 Compliance

RoHS Status: RoHS Compliant

Moisture Sensitivity Level (MSL): 1 Unlimited

REACH Status: REACH Unaffected

A100 Estimated Pricing

Qty	Low	High
1+	$0.99	$0.99

Estimated market price range - modelled values for guidance only, not live distributor offers.

A100 Features

Built on eight NVIDIA A100 Tensor Core GPUs.
640 GB total GPU memory capacity.
600 GB/s GPU-to-GPU bandwidth via NVSwitch.
2 TB of system memory included.
High-speed networking up to 200 Gb/s.

A100 Applications

Deep learning model training with large batch sizes.
High-performance computing (HPC) simulations and analysis.
AI inference at scale for enterprise workloads.
Data analytics and scientific computing applications.
GPU-accelerated virtualization and cloud services.

A100 FAQs

7 frequently asked questions generated from this part’s specifications.

What is the input voltage requirement for the NVIDIA A100 system?

The system requires 200-240 volts AC input per power supply.

What is the form factor of the NVIDIA A100 system?

It is a 6U rackmount form factor.

How many GPUs are in the NVIDIA A100 system?

It contains 8 NVIDIA A100 GPUs.

What is the total GPU memory of this system?

Total GPU memory is 640 GB.

How much system memory does the A100 system have?

It has 2 TB of system memory in its factory configuration.

What storage options are available?

OS storage is 1.92 TB NVMe M.2 SSDs in RAID 1 array; data cache storage is 3.84 TB NVMe U.2 SED in RAID 0 array (optional 7.68 TB drives).

What network interfaces are supported?

Cluster network: single-port InfiniBand or Ethernet up to 200 Gbps; storage network: dual-port Ethernet up to 200GbE, with backward compatibility.