NVIDIA DGX Spark

A Grace Blackwell AI supercomputer on your desk — 1 PFLOPS AI performance, 128GB unified memory, 6,144 CUDA cores in a compact 1.2kg form factor.

🚀 Express Shipping Available Across Europe & MENA

  • Check Mark Full Insurance on All Shipments
  • Check Mark Tracked Delivery & Real-Time Updates
GUARANTEED SAFE CHECKOUT
  • Stripe
  • Visa Card
  • MasterCard
  • American Express
  • Discover Card

Overview

The NVIDIA DGX Spark is a personal AI supercomputer powered by the NVIDIA GB10 Grace Blackwell Superchip. Delivering 1 petaFLOP of AI performance in a form factor that fits in the palm of your hand, it enables developers, researchers, and data scientists to run inference on models with up to 200 billion parameters and fine-tune models up to 70 billion parameters — all locally on their desk.

With 128GB of unified LPDDR5x memory shared between the Blackwell GPU and 20-core Arm CPU via NVLink-C2C, DGX Spark eliminates the memory bottlenecks that limit AI development on traditional desktops. It comes preloaded with NVIDIA DGX OS and the full NVIDIA AI software stack, so you can start building AI from the moment you unbox it.

Key Features

  • 1 PFLOPS AI Performance: GB10 Grace Blackwell Superchip with 5th Gen Tensor Cores
  • 128GB Unified Memory: LPDDR5x with 273 GB/s bandwidth, shared CPU+GPU coherent
  • 6,144 CUDA Cores: Blackwell GPU with 4th Gen RT Cores
  • 20-Core Arm CPU: 10x Cortex-X925 + 10x Cortex-A725 for balanced compute
  • ConnectX-7 200GbE: High-speed networking for multi-node and cloud connectivity
  • Ultra-Compact: 150 x 150 x 50.5 mm, 1.2 kg — sits on any desk
  • Preloaded Software: DGX OS with full NVIDIA AI stack ready out of the box

Technical Specifications

Specification Details
Processor NVIDIA GB10 Grace Blackwell Superchip
GPU NVIDIA Blackwell, 6,144 CUDA cores
CPU 20-core Arm (10x Cortex-X925 + 10x Cortex-A725)
Memory 128 GB LPDDR5x unified (273 GB/s)
Interconnect NVLink-C2C (5x PCIe Gen5 bandwidth)
AI Performance 1 PFLOPS
Storage 1 TB or 4 TB NVMe M.2 (self-encrypting)
Networking ConnectX-7 (2x QSFP 200GbE), 10GbE RJ45, Wi-Fi 7, Bluetooth 5.4
Display 1x HDMI 2.1a
USB 4x USB Type-C
Power 240W external power supply
Dimensions 150 x 150 x 50.5 mm
Weight 1.2 kg

Ideal Use Cases

  • Personal AI development — run and fine-tune LLMs up to 70B parameters locally
  • AI research and prototyping with full data privacy
  • Data science workflows with large datasets on-device
  • AI education — university labs and student development
  • Enterprise AI developer workstations with secure local compute

Why Choose This Product?

DGX Spark puts a genuine AI supercomputer on every developer’s desk. With 128GB of unified memory and 1 PFLOPS of AI performance, it runs models that previously required cloud infrastructure — with zero latency, full data privacy, and no recurring cloud costs. For organizations building AI teams, DGX Spark is the most productive way to equip every developer.

Interested? Contact us for pricing, bulk orders for teams, and enterprise deployment support.

Reviews

There are no reviews yet.

Be the first to review “NVIDIA DGX Spark”

Your email address will not be published. Required fields are marked *