NVIDIA DGX Spark - 4TB A GB10 Grace Blackwell AI Supercomputer on your desk

Click to open expanded view

Global Local Service

Lifetime Technical Support

NVIDIA DGX Spark Founders Edition (940-54242-0006-000) AI Server 4TB NVMe 128GB DDR5X | Personal AI Computer for Deep Learning, LLM & GPU Computing

Name: NVIDIA DGX Spark Founders Edition (940-54242-0006-000) AI Server 4TB NVMe 128GB DDR5X | Personal AI Computer for Deep Learning, LLM & GPU Computing
Brand: Naddod
SKU: 103029
Price: 4499.00 USD
Availability: InStock
Rating: 5.0 (18 reviews)

#103029

Model: DGX Spark (940-54242-0006-000)

Sold: 0

In Stock: 45

Ships within 48 hours

DGX Spark Types:

DGX Spark

DGX Spark Bundle

DGX Spark

DGX Spark Bundle

Item Spotlights

DGX Spark Interconnect Cable — Q56-200G-CU0-5
NVIDIA GB10 Grace Blackwell superchip
1 PFLOPS of FP4 AI performance
128GB of coherent, unified system memory
ConnectX-7 Smart NIC
4 TB NVME.M2 with self-encryption
150mm L x 150mm W x 50.5mm H

NVIDIA DGX Spark Founders Edition (940-54242-0006-000) AI Server 4TB NVMe 128GB DDR5X | Personal AI Computer for Deep Learning, LLM & GPU Computing

Free Worldwide Shipping on Orders over $ 99

Specifications

Applications

Product Highlights

Videos

Questions & Answers

Customer Reviews

Resources

Specifications

Applications

Product Highlights

Videos

Questions & Answers

Customer Reviews

Resources

Description

NVIDIA DGX Spark AI Desktop Platform – Grace-Blackwell GB10, 128 GB LPDDR5x, 4 TB NVMe, 1 PFLOP Tensor FP4

Powered by the NVIDIA GB10 Grace Blackwell Superchip, NVIDIA DGX™ Spark delivers one petaFLOP of AI performance in a power-efficient, compact form factor. With the NVIDIA AI software stack preinstalled and 128 GB of memory, developers can prototype, fine-tune, and inference the latest generation of reasoning AI models from DeepSeek, Meta, NVIDIA, Google, Qwen and others with up to 200 billion parameters locally.

Specifications

Architecture

NVIDIA Grace Blackwell

GPU

NVIDIA Blackwell Architecture

CPU

20 core Arm,10 Cortex-X925 + 10 Cortex A725 Arm

CUDA Cores

NVIDIA Blackwell Generation

Tensor Cores

5th Generation

RT Cores

4th Generation

Tensor Performance

1000 AI TOPS

System Memory

128GB LPDDR5x, coherent unified system memory

Memory Interface

256-bit

Memory Bandwidth

273 GB/s

Storage

4 TB NVME.M2 with self-encryption

USB

4x USB Type C

Ethernet

10 GbE, 1x RJ-45 connector

NIC

ConnectX-7 Smart NIC

Wi-Fi

Wi-Fi 7

Bluetooth

BT 5.4 w/LE

Audio output

HDMI multichannel audio output

Power Consumption

240 W

Display Connectors

1x HDMI 2.1a

NVENC | NVDEC

1x | 1x

NVIDIA DGX™ Base OS, Ubuntu Linux

System Dimensions

150 mm L x 150 mm W x 50.5 mm H

System Weight

1.2 kg

Applications

Product Highlights

Why NADDOD

Module Availability

Worry-Free - Rigorous authentic testing!
Multi-version Device Compatibility

Multi-Version - Easy adaptation for each module!
High-Performance

Actual Test Data-Better results intuitively!

Our Approach

Test scene

NADDOD stringently selects a series of items in a professional testing environment to ensure the perfect blend of product performance and reliability. Including hardware compatibility (plugging, unplugging, rebooting), software compatibility (connectivity, parameters) and performance testing (DDM, BER, stability).

Test result

NVIDIA DGX Spark

Linux version: 6.11.0-1014

Connectivity

DGX Spark Side Testing

Connectivity Testing

The DAC cable can connect to two DGX Spark units normally, and their port states are up.
BER

DGX Spark Side Testing

BER Testing

Test REAL-TIME BER to ensure the accuracy of data transmission under high load conditions.
Bandwidth

DGX Spark Side Testing

Bandwidth Testing

NADDOD's 0.5m 200G QSFP56 DAC cable enables DGX Spark to achieve stable, high-speed dual-unit communication, ideal for demanding computing scenarios.

Videos

DGX Spark and Cable Connection & Testing | NADDOD

Questions & Answers

Ask a Question

What are the application scenarios of NVIDIA DGX Spark?

DGX Spark is designed specifically for AI development workloads such as prototyping, model fine-tuning, inference, and data science. The DGX Spark system can be configured in standalone mode, directly connected to a monitor, keyboard, and mouse. It can also be deployed in headless mode, serving as a network-connected computing device without direct connections to a monitor, mouse, or keyboard.

What are the DGX Spark's power requirements?

The DGX Spark can be powered via a standard wall outlet.

What cable is recommended for stacking/clustering two systems?

A certified QSFP cable—such as Q56-200G-CU0-5 —is supported for connecting two DGX Spark systems.

How many DAC cables are recommended for connecting two DGX Spark units?

It is recommended to configure 2 DAC cables, supporting flexible usage: you can plug in one for daily operation and keep the other as a backup, or plug in both simultaneously for connectivity. Both setups are supported, offering versatile connectivity options.

When connecting two DGX Spark units with one or two 0.5m 200G QSFP56 DAC cables, is an Ethernet cable still required?

Yes, you'll need an Ethernet cable or a Wi-Fi connection. DAC cables are dedicated to data transmission between devices, while the two DGX Spark units rely on an out-of-band management network for administrative communication. This network offers two options: a wired connection (Ethernet via RJ45 port, for better stability) or Wi-Fi (for easier use), so you can choose based on your needs.

What operating system (OS) does DGX Spark run?

DGX Spark runs DGX OS based on Ubuntu, which is the same operating system distribution as that running on the DGX platform.

What is the GB10 Grace Blackwell Super Chip?

The GB10 Super Chip is a system-on-chip (SoC) based on the Grace Blackwell architecture, featuring 128GB of unified system memory with unified addressing. It delivers up to 1 petaFLOP of AI performance at FP4 precision, integrating an NVIDIA Blackwell GPU equipped with the latest 5th-Generation Tensor Cores and RT Cores, as well as a high-performance Grace CPU with 20 Arm cores.

What are the advantages of the GB10 Architecture?

The GB10 Architecture enables enterprises and researchers to prototype, tune, and test models on local DGX Spark systems, then deploy them in data centers or the cloud. This supports efficient fine-tuning and inference workloads while maintaining consistency between development and deployment environments.

What memory and storage do two DGX Spark systems configured as a cluster provide?

Two DGX Spark systems can be configured as a cluster, offering 256GB of combined memory and up to 8TB of total storage.

Customer Reviews

5.0

5 Reviews

5 Star
100%
4 Star
0%
3 Star
0%
2 Star
0%
1 Star
0%

All Reviews (5)

k***8
Verified Purchase
k***8
Verified Purchase
5.0
I bought four DGX Sparks and four 0.5m 200G Q56 DACs. For daily use, I connect each pair of units with either one or two DACs—both work great. The remaining DACs serve as spares for redundancy. Whether using a single DAC or two, the setup runs perfectly stable with smooth, consistent performance. If a DAC ever fails, I can swap in a replacement right away. I'm really satisfied with how reliable and hassle-free the whole setup is.
Helpful (7)Comments
m***l
Verified Purchase
m***l
Verified Purchase
5.0
I bought four DGX Sparks and wanted to connect them to my existing NVIDIA SN3700 switch. I also bought the four 200G Q56 DACs recommended for this connection later. Connecting the four DGX Sparks to the switch with these cables works well. Everything runs stably, and the performance is good. I'm happy with how it's working.
Helpful (1)Comments
h***6
Verified Purchase
h***6
Verified Purchase
5.0
Grabbed my DGX Spark and the DAC cable for linking two units, and wow, shipping was pretty speedy! The DGX Spark is surprisingly tiny—sits on my desk without hogging space, and it's easy to carry around too. The DAC cable plugs into the ports perfectly. Both are a good choice!
Helpful (8)Comments
r***n
Verified Purchase
r***n
Verified Purchase
5.0
Got my NVIDIA DGX Spark—great! Runs smooth for AI and data tasks. Super worth it!
Helpful (7)Comments
q***t
Verified Purchase
q***t
Verified Purchase
5.0
I got two DGX Spark units and a DAC cable for connecting them, and man, the shipping was super fast! Ordered it and got everything in no time. Both the DGX Spark and the DAC work like a dream—super reliable and exactly what I want.
Helpful (2)Comments