NVIDIA DGX Spark
Accelerate All AI Workloads
NVIDIA DGX Spark Personal AI Supercomputer delivers a petaflop of AI performance from a power-efficient, compact form factor. With the preinstalled NVIDIA AI software stack, 1000 TOPS of performance, and 128GB of memory, developers can prototype, fine-tune, and inference large AI models of up to 200 billion parameters locally and seamlessly deploy to the data centre or cloud.
NVIDIA DGX SPARK ASUS ASCENT GX10 1TB AI Computer
NVIDIA DGX™ OS (Pre-installed NVIDIA AI Software Stack) ARM v9.2-A CPU (GB10) NVIDIA Blackwell GPU (GB10, integrated) 128GB LPDDR5x Coherent Unified Memory 1TB NVMe M.2 2242 PCIe 4.0
NVIDIA DGX Spark MSI Edgexpert AI Computer
NVIDIA DGX™ OS (Pre-installed NVIDIA AI Software Stack) 20 core Arm, 10 Cortex-X925 + 10 Cortex-A725 Arm NVIDIA GB10 Grace Blackwell Superchip 128GB LPDDR5x Unified System Memory 4TB NVMe M.2 PCIe 5.0 with Self-Encryption
0 Review
Estimated dispatch date: Monday, 12/10/2026
200B Parameter Models on Your Desk
The increasing size and complexity of generative AI models is making development efforts on local systems challenging. Prototyping, tuning, and inferencing large models locally requires large amounts of memory and significant compute performance. As enterprises, software providers, government agencies, startups, and researchers staff up AI efforts, the need for AI compute resources continues to grow.
NVIDIA DGX™ Spark is part of a new class of computers designed from the ground up to build and run AI. Powered by the NVIDIA GB10 Grace Blackwell Superchip and based on the NVIDIA Grace Blackwell architecture, NVIDIA DGX Spark delivers up to 1000 TOPS of AI performance to power large AI workloads.
Experience up to 1000 AI TOPS of AI performance at FP4 precision with the NVIDIA Grace Blackwell architecture.
Run AI development and testing workloads with AI models up to 200 billion parameters at your desktop with large, unified system memory.
High-performance NVIDIA Connect-X networking enables connecting two NVIDIA DGX Spark systems together to work with AI models up to 405 billion parameters.
Use a full-stack solution for generative AI workloads, including tools, frameworks, libraries, and pretrained models.
Connect Two DGX Systems
Upgrade easily by Connecting two DGX Spark systems together via ConnectX-7 and LinkX cable to work with even larger models of up to 405 billion parameters utilising the full 256GB of unified memory for better than multi-GPU workstation performance*
*Assuming 4-bit precision and workloads specific to high memory usage models. See comparison table for more information.
Develop Locally, Deploy Anywhere
DGX Spark comes with all the software AI developers need to get up running AI workloads right away. Users can install their favourite tools, and frameworks or take advantage of the latest NVIDIA workflows, models, microservices and tools such as NVIDIA NIM, AI Blueprints or AI Workbench.
Frequently Asked Questions
What is the NVIDIA DGX Spark?
The NVIDIA DGX Spark is an AI supercomputer built to sit on a desk for personal use, rather than in a data centre. Built around the NVIDIA GB10 Grace Blackwell Superchip, the DGX Spark delivers up to 1,000 AI TOPS at FP4 precision alongside 128GB of unified memory, which lets developers prototype, fine-tune and run AI models of up to 200 billion parameters locally. It ships with NVIDIA's AI software stack preinstalled, so work developed on it deploys to a data centre or cloud without rewriting.
What AI models can a DGX Spark run?
The DGX Spark's 128GB of unified memory allows it to locally run and fine-tune models of up to 200 billion parameters. Connecting two units together via ConnectX-7 raises that to models of up to 405 billion parameters, using the full 256GB of combined unified memory. These figures assume 4-bit precision, so higher-precision work will fit smaller models into the same memory.
What is the NVIDIA GB10 Grace Blackwell Superchip?
The GB10 is the processor at the core of the DGX Spark, combining NVIDIA's Grace CPU architecture with Blackwell GPU architecture on a single package. That pairing is what allows the system to share 128GB of unified memory between the processor and GPU, rather than splitting it as a conventional workstation does. It delivers up to 1,000 AI TOPS at FP4 precision in a desktop form factor.
Who is the DGX Spark for?
The DGX Spark is best suited for professional developers, researchers and teams building AI applications who need local compute rather than cloud instances. That usually means organisations handling data unable to leave their premises, teams wanting predictable costs rather than hourly billing, or developers who need a system available immediately instead of queued. Work prototyped on a DGX Spark deploys to a data centre or cloud without modification, which makes it a development machine rather than a replacement for production infrastructure.
Where can I buy the NVIDIA DGX Spark in the UK?
CyberPowerPC UK supplies the NVIDIA DGX Spark to its customers, UK customers have the additional benefit of free standard delivery to UK mainland addresses and lifetime support from a UK-based team. Business and education customers can be quoted directly, including for multiple units.