NVIDIA DGX Spark: AI Supercomputer on Your Desk

7 min read Original article ↗

The power of NVIDIA Grace Blackwell on your desk.

Where to Buy

Find your best purchase option on the NVIDIA Marketplace.

Previously Project DIGITS

Overview

Spark Something Big

Powered by the NVIDIA GB10 Grace Blackwell Superchip, NVIDIA DGX Spark™ delivers up to one petaFLOP1 of FP4 AI performance in a power-efficient, compact form factor. With a preinstalled NVIDIA AI software stack and 128 GB of memory, developers can prototype, fine-tune, and deploy the latest reasoning AI models. Additionally, NVIDIA NemoClaw, part of the NVIDIA Agent Toolkit is an open source agent development platform for building, evaluating, and optimizing safer, long-running autonomous agents directly from the desktop.

1 Theoretical FP4 TOPS using the sparsity feature. 

NVIDIA Announces NemoClaw for OpenClaw Community

NemoClaw adds security and privacy to run secure, always-on AI assistants on NVIDIA RTX PCs, DGX Station, and DGX Spark.

DGX Spark and DGX Station with NVIDIA NemoClaw for Autonomous Agents

The ultimate platform for locally developing and deploying secure, supercomputing-intelligent, always-on AI agents.

Run Autonomous Agents More Safely

NVIDIA NemoClaw is an open source stack that adds privacy and security controls to OpenClaw. With one command, anyone can run always-on, self-evolving agents anywhere.

Playbooks for Everyone

From novice to expert, leverage our curated playbooks to get off to a running start on various projects. Playbooks contain step-by-step recipes to demonstrate the art of the possible.

Getting Started With Your DGX Spark

Here are some resources that can help you get up and running fast.

Features

NVIDIA GPU, CPU, Networking, and AI Software Technologies

NVIDIA GB10 Superchip

Experience up to  1 petaFLOP  of AI performance at FP4 precision with the NVIDIA Grace Blackwell architecture.

128 GB of Coherent Unified System Memory

Run AI development and testing workloads with AI models up to 200 billion parameters at your desktop with a large, unified system memory.

NVIDIA ConnectX Networking

High-performance NVIDIA ConnectX™ networking enables the connection of two NVIDIA DGX Spark systems to work with AI models of up to 405 billion parameters.

NVIDIA AI Software Stack

Utilize a full-stack solution for generative AI workloads, encompassing NVIDIA tools, frameworks, libraries, and pre-trained models including NVIDIA NIM.

Workloads

Accelerate All AI Workloads

Delivering the power of Grace Blackwell  in a desktop-friendly size, NVIDIA DGX Spark is ideal for AI developer, researcher, and data scientist workloads.

Prototyping

Develop, test, and validate AI models and applications.

With the NVIDIA AI software stack, NVIDIA DGX Spark provides a platform for developers to create, test, and validate AI models, as well as build AI agents, AI-augmented applications, and solutions. For final tuning or deployment, conveniently evaluate the work for eventual migration to the NVIDIA DGX cloud or other NVIDIA-accelerated data centers or cloud infrastructures.

Prototyping

Fine-Tuning

Fine-tune AI models up to 70 billion parameters.

Improve the performance of pre-trained models by fine-tuning on NVIDIA DGX Spark. With 128GB of unified system memory, fine-tune models up to 70 billion parameters to customize AI models and solutions for specific needs and use cases.

Fine-Tuning

Inference

Test, validate, and inference with AI models up to 200 billion parameters.

Fifth-generation Tensor Cores with support for FP4 deliver up to 1 petaFLOP of AI computing performance, combined with 128GB of system memory, accelerate inference of state-of-the-art AI models to test, validate and deploy from your NVIDIA DGX Spark.

Inference

Data Science

High-performance data science at your desk.

NVIDIA DGX Spark’s combination of 128GB of unified memory and 1 petaFLOP of parallel throughput maximizes performance of large, computationally complex data analytics and machine learning workflows at your desk.

Data Science

Edge Applications

Develop edge applications with NVIDIA AI frameworks, including NVIDIA Isaac™, Metropolis, and many others.

NVIDIA DGX Spark offers an exceptional platform for developing robotics, smart city, and computer vision solutions. NVIDIA frameworks include Isaac, Metropolis, and Holoscan, enabling developers to leverage the power of NVIDIA DGX Spark to develop edge applications quickly.

Edge Applications

NVIDIA AI Enterprise—DGX Spark

A cloud-native suite of software tools, libraries, and frameworks accelerating production-grade AI development. Combines enterprise-grade security, optimized performance, and enterprise-level support to streamline prototype-to-production for next-gen agentic AI.

DGX Spark AI Software

DGX Spark includes the NVIDIA AI software stack preinstalled with support for the NVIDIA AI software ecosystem to get AI projects up and running quickly.

Explore NVIDIA DGX Spark

March 16, 2026

Scaling Autonomous AI Agents and Workloads with NVIDIA DGX Spark

Learn why NVIDIA DGX Spark is an ideal desktop platform for autonomous AI.

February 12, 2026

NVIDIA DGX Spark Powers Big Projects in Higher Education

The desktop supercomputer is sparking innovation across research fields, from campus labs to the South Pole.

January 5, 2026

NVIDIA DGX Spark Gains 2x Performance and Open AI Model Support

The latest DGX Spark software release delivers performance uplift across models and workflows.

October 13, 2025

NVIDIA DGX Spark Arrives for World’s AI Developers

NVIDIA and its partners are shipping DGX Spark, the world’s smallest AI supercomputer, delivering NVIDIA's AI stack in a compact desktop form factor.

February 20, 2026

Meet the AI Photo Booth: DGX Spark + Reachy Mini at CES

At CES 2026, the NVIDIA DGX Spark powers the Reachy Mini robot in an interactive photo booth with Pollen Robotics (Hugging Face).

January 5, 2026

Build Your Own AI Assistant With Hugging Face on NVIDIA DGX Spark

NVIDIA and Hugging Face are bringing AI agents to life.

October 15, 2025

Sparking Something Big: NVIDIA DGX Spark Has Arrived

It’s here! Go behind the scenes and see the delivery of NVIDIA DGX Sparks to top developers, researchers, creative studios, and roboticists.

NVIDIA DGX Spark Livestreams

Watch live developer deep dives on the new NVIDIA DGX Spark desktop AI supercomputer.

DGX Spark/GB10 User Forum

Join the DGX Spark/GB10 user community. Learn from each other, get support from experts, and be inspired to create the next great AI.

NVIDIA Inception

Evolve your startup with go-to-market support, technical expertise, training, and funding opportunities.

Partners

NVIDIA DGX Spark

Get DGX Spark through our authorized channel and retail partners.

NVIDIA GB10 Superchip Powered Systems

Discover NVIDIA GB10 Grace Blackwell Superchip–powered systems from our preferred OEM partners.

NVIDIA DGX Spark Specifications

1 Theoretical FP4 TOPS using the sparsity feature.

2 TDP: Thermal Design Power of the GB10 chip including CPU and GPU

Declared noise emission values in accordance with ECMA-109, June 2025

Product name NVIDIA DGX Spark
940-54242-0000
Product description NVIDIA GB10 Grace Blackwell
Superchip
20-core Arm: 10 Cortex-X925 + 10
Cortex-A725 Arm,
128 GB LPDDR5x
4 TB NVME.M2
Quantities declared Operating Modedd) Idle
Declared mean A-weighted sound power level a), LWA,m (dB) 35 19
Declared mean A-weighted emission sound pressure level b), LpA,m (dB) 29 13
Statistical adder for verification c), Kv (dB) 3 3

a) The declared mean A-weighted sound power level, Lwa,m is computed as the arithmetic average of the measured A-weighted sound power levels.

b) The declared mean A-weighted emmison sound pressure level, LpA,m is computed as the arithmetic average of the defined Operator postiion (C.4b)

c) The statistical adder for verification, KV is a factor to be added to the declared mean A-weighted sound power level, LWA,m, such that there will be a 95 % probability of acceptance, when using the verification procedures of ECMA-109, if no more than 6,5 % of the equipment in batch, has A-weighted sound power levels greater than (LWA,m + Kv).

d) Operating mode, max GPU stress in 25°C ambient.

Note 1 The quantity, LWA,c (formerly called LWAd) can be computed from the sum of LWA,m and KV.

Note 2 All measurements made in conformance with ECMA-74 and declared with ECMA-109

Note 3 dB is the abbreviation for decibles

Sign up To Be Notified on Availability

Welcome back. Not you? Log Out