NVIDIA DGX™ Spark

The NVIDIA DGX™ Spark is a compact desktop AI workstation designed to bring powerful AI performance directly to your desk. Built for developers, researchers, and AI engineers, it supports building, testing, fine-tuning, and running advanced AI models locally before deploying them to data centers or cloud platforms.

Powered by the NVIDIA GB10 Grace Blackwell Superchip, DGX Spark delivers up to 1 PFLOP (1000 TOPS) of AI performance in a compact form factor. With 128 GB of unified system memory, it handles large AI workloads and models with up to 200 billion parameters for machine learning, deep learning, generative AI, and reasoning tasks.

The system integrates the complete NVIDIA AI software stack, including tools and frameworks such as CUDA, CUDA-X, PyTorch, and TensorFlow. NVIDIA DGX OS provides a consistent development environment and allows easy movement of workloads between local systems, data centers, and cloud platforms.

DGX Spark is designed to be space-efficient and power-efficient, making it suitable for AI labs, universities, enterprise teams, and research environments in the UAE. Its compact design fits standard workspaces while still delivering enterprise-level AI performance.

Built on NVIDIA GB10 Grace Blackwell Superchip

The NVIDIA GB10 Grace Blackwell Superchip combines high-performance CPU cores with a Blackwell GPU to support demanding AI workloads. This design helps maintain stable performance for training, fine-tuning, and inference tasks.

Key Features

• Up to 1 PFLOP (1000 TOPS) of AI performance (FP4)
• 128 GB unified system memory for large AI models
• Supports AI models up to 200B parameters
• Complete NVIDIA AI software stack for AI development
• 4 TB NVMe storage for fast local data access
• High-speed networking support for data-intensive workloads
• Compact desktop form factor for office and lab environments

Software & Development Environment

NVIDIA DGX Spark ships with NVIDIA DGX OS and supports popular AI frameworks, allowing teams to start development quickly. The consistent software stack ensures smooth scaling from local development to larger DGX systems or cloud-based deployments.